Skip to main content

Full text of "The Founders Dilemmas (howentrepreneur.com) By Noam Wasserman"

See other formats


THE—.* 1 

FOUNDERS 



DILEMMAS 



/ ANTICIPATING AND 

AVOIDING THE 
PITFALLS 
Y THAT CAN 

■ SINK 

A 

STARTUP 




NOAM WASSERMAN 



THE FOUNDER'S DILEMMAS 



THE KAUFFMAN FOUNDATION SERIES ON INNOVATION AND ENTREPRENEURSHIP 



Boulevard of Broken Dreams: Why Public Efforts to Boost Entrepreneurs hip and Venture Capital Have Failed — and What to 
Do about It, by Josh Lemer 

The Invention of Enterprise: Entrepreneurship from Ancient Mesopotamia to Modem Times , edited by David S. Landes, Joel 
Mokyr, and William J. Baumol 

The Venturesome Economy: How Innovation Sustains Prosperity in a More Connected World, by Amar Bhide 

The Microtheory of Innovative Entrepreneurship, by William J. Baumol 

Solomon 's Knot: How Law Can End the Poverty of Nations, by Robert D. Cooter and Hans-Bemd Schafer 



THE 

FOUNDER'S 



DILEMMAS 



ANTICIPATING AND 
AVOIDING THE PITFALLS 
THAT CAN SINK 
A STARTUP 

NOAM WASSERMAN 



PRINCETON UNIVERSITY PRESS PRINCETON & OXFORD 



Copyright © 2012 by Princeton University Press 

Published by Princeton University Press, 

41 William Street, Princeton, New Jersey 08540 

In the United Kingdom: Princeton University Press, 
6 Oxford Street, Woodstock, Oxfordshire OX20 ITW 

pr e ss . princ e ton. e du 

All Rights Reserved 

Library of Congress Cataloging-in-Publication Data 

Wasserman, Noam, 1969- 
The founder's dilemmas : anticipating and avoiding the pitfalls that can sink a startup / Noam Wasserman. 

p. cm. — (The Kauffinan foundation series on innovation and entrepreneurship) 
Includes bibliographical references and index. 
ISBN 978-0-691-14913-4 (hardcover) 

1. New business enterprises — Management. 2. Entrepreneurship.~I. Title. 

HD62.5.W375 2012 

658.1'l^c23 

2011037954 

British Library Cataloging-in-PubKcation Data is available 

Published in collaboration with the Ewing Marion Kauffinan Foundation and the Berkley Center for Entrepreneurial Studies, New York 

University 

This book has been composed in Sabon and Trade Gothic 

Printed on acid-free paper, co 

Printed in the United States of America 



10 987654321 



CONTENTS 



List of Illustrations 

Part I: Introduction and Pre-founding 

CHAPTER ONE 

Introduction 

CHAPTER TWO 

Career Dilemmas 

Part n: Founding Team Dilemmas 

CHAPTER THREE 

The Solo-versus-Team Dilemma 

CHAPTER FOUR 

Relationship Dilemmas: Flocking Together and Playing with Fire 

CHAPTER FIVE 

Role Dilemmas: Positions and Decision Making 

CHAPTER SIX 

Reward Dilemmas: Equity Splits and Cash Compensation 

CHAPTER SEVEN 

The Three Rs System: Alignment and Equilibrium 

Part ni: Beyond the Founding Team: Hires and Investors 

CHAPTER EIGHT 

Hiring Dilemmas: The Right Hires at the Right Time 

CHAPTER NINE 

Investor Dilemmas: Adding Value, Adding Risks 

CHAPTER TEN 

Failure, Success, and Founder-CEO Succession 
Part rV: Conclusion 

CHAPTER ELEVEN 

Wealth- versus-Control Dilemmas 



Acknowledgments 

Appendix A: Quantitative Data 

Appendix B: Summary of Startups and People 

Notes 

Bibliography 
Index 



ILLUSTRATIONS 



1 . 1 Sequence of Founding Dilemmas 

1.2 Wealth-versus-Control Dilemmas 

1.3 Principal Case Studies 

2. 1 Factors Affecting the "When to Found" Decision 

2.2 Developments That Release Each Type of Handcuff 

2.3 Three Big Career Questions 

2.4 Career, Personal, and Market Circumstances 

2.5 Obstacles and Potential Solutions for Gray-Area Potential Founders 

3. 1 Sizes of Founding Teams for Technology and Life Sciences Startups 

3.2 Central "Solo versus Team" Questions 

3.3 When to Solo-Found versus Build a Founding Team 

3.4 Advantages and Disadvantages of Going Solo versus Building a Founding Team 

4. 1 Relationship Dilemmas, in the Context of the Broader Set of Founding Dilemmas 

4.2 Effects of Homogeneity within the Founding Team 

4.3 The Playing-with-Fire Gap 

5. 1 Degree of Top-Heaviness in Two-Founder and Three-Founder Teams, at Time of Founding 

5.2 Percentages of "Idea" and "Non-idea" Founders Who Received Formal Titles at Time of Founding 

5.3 Strengths and Weaknesses of Overlapping Roles versus Division of Labor 

5.4 Egalitarian versus Hierarchical Decision Making 

6. 1 Reasons to Split Earlier versus Later 

6.2 Founders' Capital Contributions 

6.3 Factors for Determining Equity Splits 

6.4 Difference in Equity Stakes between Founder with Largest Stake and Founder with Smallest Stake 

6.5 Time Spent Negotiating the Split versus Formality of the Agreement 

6.6 UpDown's Template for Splitting Equity 

7. 1 The Three Rs System 

7.2 The Linkage between Prior Relationships, Equity Splits, and Team Stability 

III. 1 Beyond-the-Team Dilemmas, in the Context of the Broader Set of Founding Dilemmas 

8. 1 Sources of Executive Hires in Startups Led by Founder-CEOs 

8.2 Changes in Executive Teams as Startups Mature 

8.3 Changes in Structural Leverage as the Startup Grows 

8.4 Seniority of Hires, by Maturity of Startup (number of rounds of financing raised) 

8.5 C-levelNonfounder Salaries, 2000-2009 

8.6 Bonus as Percentage of Salary for Nonfounding Executives in Technology Startups, across the Full Decade 

8.7 Nonfounding Executives: Compensation and Equity Inequality 

8.8 Years of Vesting for Founder-CEOs versus Nonfounder-CEOs, across All Stages of Financing 

8.9 Evolution of Hiring Decisions as the Startup Matures 

9. 1 Central "Investor Dilemmas" Questions for Startups That Can Attract Outside Financing 

9.2 Pros and Cons of Investor Options 

9.3 Types of Investors Participating in Each Round of Outside Financing in the CompStudy Dataset 

9.4 Sources of Leads to Initial Angel and Venture-Capital Investors 

9.5 Capital Invested and Pre-Money Valuation for Each Round of Financing 

9.6 Capitalization Tables: Average Equity Holdings of Each Major Party 

9.7 Percentage of Financing Rounds That Were Down Rounds, by Year the Round Was Raised 

10. 1 Percentage of Startups Still Being Led by the Founder-CEO 

10.2 The Founder-CEO Succession Process 

10.3 Post-Succession Positions of Replaced Founder-CEOs 
11.1 Wealth-versus-Control Dilemmas 



11.2 Rich versus King Outcomes 

11.3 Value of Founder's Equity Stake, Depending on Degree of Control Maintained 

1 1.4 A Model of Founder Decisions 

11.5 Reasons Not to Sell versus Reasons to Sell 



PARTI 

INTRODUCTION AND 
PRE-FOUNDING 



CHAPTER ONE 

INTRODUCTION 



IT'S UNFORTUNATE BUT TRUE: If entrepreneurship IS A BATTLE , most casualtics Stem from friendly 
fire or self-inflicted wounds. Some four decades ago, sociologist Arthur Stinchcombe attributed much 
of the "liability of newness" — the particularly high failure rate of new organizations — to problems 
within a startup's founding team. More recently, venture capitalists in one survey attributed 65% of 
failures within their portfolio companies to problems within the startup's management team ^ Another 
study asked investors to identify problems they thought might occur within their portfolio companies; 
a frill 61% of such problems involved issues within the team^ 

Researchers have extensively studied the failure rates highlighted by Stinchcombe. Unfortunately, 
they have focused almost entirely on external causes rather than on the more numerous internal 
problems Stinchcombe identified.* We know amazingly little about the chief perils that beset the 
entrepreneurial activity we so often acclaim as the very heart and soul of the economy. As researcher 
Amar Bhide put it, "Enfrepreneurs who start and build new businesses are more celebrated than 
studied."^ 

This book scrutinizes those all-inportant "people problems" that bedevil all founders — even solo 
founders — and their startups.* These problems often follow predictably from common dilemmas 
faced by any startup as it grows and evolves — ^what I call "founding dilemmas." One such dilemma 
recurs throughout the stages of company development: The need to negotiate a frade-off between 
wealth and control, between building financial value and maintaining a grip on the steering wheel. 
Adding frirther complexity, founders' early choices can have delayed and unexpected but significant 
effects,^ sometimes because natural inclinations such as passion, optimism, and conflict avoidance 
lead to shortsighted decisions. This book delves into the challenges faced or created by each of the 
main groups of players involved in a typical startup, beginning with the core founder and moving on 
to his or her cofounders, early hires, and investors. 

I've spent over a decade working with hundreds of founders and friture founders, and I've collected 
and analyzed data on nearly 10,000 founders in the technology and life sciences industries. This book 
will draw on my unique dataset, and we will also follow actual founders as they launch new 
conpanies and sfruggle with dilemmas identified in my research. Most centrally, we explore the 
experience of Evan Williams, a young entrepreneur who moved from rural Nebraska to the San 
Francisco area in the mid-1990s, hoping to catch the wave of the Internet boom after working on a 
failed startup back home. A self-taught Web designer and programmer, Evan recognized the potential 
of Internet applications, particularly in the exploding area of self-publishing, and created Blogger, 
one of the first and most popular blogging tools. Later, Evan developed an early podcasting idea, 
Odeo, which he believed would enable nontechnical people to produce, publish, and share audio 
content, much as Blogger had done for millions of ordinary people with the written word. 

In both startups, and then later when he went on to found and lead Twitter before leaving to work 



on yet another startup, Evan faced key dilemmas — difficult decisions at inportant forks in the road of 
his entrepreneurial journey — and took steps that would shape the startup's future, affect its value, and 
help determine his degree of control over it. ^\^th Blogger, Evan chose to cofound with his former 
girlfriend, fighting to retain for himself the CEO title and a majority of the equity. He funded the 
conpany using his own money along with money from friends, family, and angel investors, 
intentionally avoiding venture capitalists (VCs). He hired friends (and later, volunteers) to 
inexpensively develop the technology. When the dot-com boom ended and Blogger ran low on cash, 
he received an acquisition offer but refused to sell — a decision that would send his cofounder and his 
entire staff fleeing for the exits and reduce him to soliciting donations to keep Blogger alive. 

Evan eventually sold Blogger to Google and turned next to developing Odeo, partnering with an 
acquaintance who had experience in online audio. Using some of his proceeds from Blogger, Evan 
seeded Odeo and let his cofounder take the CEO role. As Evan realized the huge potential for 
podcasting, he took over as CEO and raised $5 million from VCs, who stepped in to help make 
significant decisions. With the VC money, Evan hired an experienced and expensive leadership team, 
hoping to develop Odeo quickly enough to stay ahead of foreseeable and formidable conpetitors such 
as Apple and Yahoo. 

Evan holds particular interest for us because his very different approaches to founding and running 
Blogger and Odeo highlight the wide variety of options available to founders. Evan's decisions may 
seem wildly inconsistent: He takes VC funding in one case and vehemently resists it in another. He 
hires his pals for next to nothing in one case and pays big bucks for the pros in another. He battles his 
own ex-girlfriend for the CEO role and finally squeezes her out of the company but then readily turns 
the reins over to a mere acquaintance in his second startup. Going deep into Evan's story, we will 
explore both the underlying consistency of his decisions and the powerful motivations and situational 
factors at work, an exercise that will enable current or would-be founders and others involved in 
startups to unlock the mystery of their own motivations and dilemmas as they enter the battle that is 
entrepreneur ship . 

CORE CONCEPTS AND ARGUMENTS 

Founding a startup can seem like a fragmented, even chaotic, way of life. Perhaps no business pursuit 
is messier than creating an organization from scratch. Founders themselves are an extremely varied 
group, and academic research on entrepreneurs is fragmented, with different researchers looking at 
different stages of the founding process, turning different disciplinary or functional lenses on the 
issues they study, and developing findings with little apparent consistency. Meant for entrepreneurial 
scholars, educators, and mentors of entrepreneurs as well as entrepreneurs themselves. The 
Founder's Dilemmas highlights the consistent, coherent, and systematic patterns and trade-offs 
involved in the most critical founding decisions, from the dreamer's decision of whether or not to 
found in the first place through the founder's decision to exit the conpany he or she has created. 

Before we can get started, we must define some core concepts and arguments. I focus the 
discussion on the founding of high-potential startups; that is, startups (often technology- or science- 
based) with the potential to become large and valuable, even though their founders may subsequently 
make decisions that limit their growth. Where necessary, I integrate relevant findings from research 
on high-potential startups with the larger literature on founding small businesses, careful to note 
differences between the founding of high-potential startups and the founding of small businesses that 



are designed to remain small and owner-operated.* 

Founders, as I speak of them, are individuals who start new organizations to pursue opportunities, 
as scholar Howard Stevenson put it, "without regard to the resources they currently control."^ They 
make the early decisions that shape the startup and its growth, an influence that begins even before the 
founding itself and that can extend through all stages of the startup's development.* This book's 
central message is that these founding decisions need to be made by design, not by default. Each 
decision requires the founder to assess multiple options; there are more critical decisions — and more 
options within each decision — than many founders realize. Often, the "righf decision is by no means 
obvious, and may even be counterintuitive. In addition, it can come with a burdensome cost, making 
the decision gut- wrenching and requiring the founder to face stark trade-offs. Meanwhile, the most 
common decisions — cofounding with fi-iends, splitting equity equally among cofounders, etc. — are 
often the most fraught with peril. That's why I refer to the decisions discussed in this book as 
dilemmas. Founders who feel they have gained everything and lost nothing by an early decision that 
felt like a no-brainer may be in for a nasty surprise later on, when they find out what trade-off they 
should actually have been considering. 

The major parts of The Founder's Dilemmas explore the dilemmas founders face by progressively 
introducing new players whose involvement in a startup provokes difficult decisions that significantly 
affect the startup's direction and outcomes. We begin with the core founder, then add cofounders, and 
end with hires and investors. In each unfolding part, we examine the inpact of these players on the 
startup's outcomes — most centrally, on the stability of the founding team, the valuation of the startup, 
and the founders' abilities to keep control of the board of directors and the CEO position. An 
appendix to Chapter 1 1 also delves into exit dilemmas, which come long after founding but are faced 
by founders lucky enough to make it to that point, and which involve many of the same underlying 
factors. 

The primary founding dilemmas we'll explore, and the questions a potential founder should ask 
about them, break down into the following: 

1. Pre-foimding: Career Dilemmas — When in my career should I launch a startup? If I am 
passionate about an idea, but haven't accumulated the right career experiences yet, or the 
market is not yet receptive to my idea, or my personal situation is unfavorable, should I make 
the leap anyway? 

2 . Founding Team Dilemmas — Deciding to launch a startup introduces many more dilemmas 
regarding the startup's founders. 

a. The Solo-versus-Team Dilemma — Should I launch the business myself or try to attract 
cofounders? 

b. Relationship Dilemmas — Whom should I try to attract as cofounders: Friends? Family? 
Acquaintances? Strangers? Prior coworkers? 

c . Role Dilemmas — What positions should each of us take within the startup? Which 
decisions can we make alone, and which should we make as a team? How should we 
make those decisions? 

d. Reward Dilemmas — How should we divide equity and other financial rewards among 
the founding team? 



3 . Beyond the Founding Team — Both the startup's growth and lingering gaps in the founding 



team's abilities or resources often require founders to consider adding nonfounders and their 
resources. This introduces further dilemmas. 

a . Hiring Dilemmas — What types of people should I hire at different stages of growth? 

What challenges will my early hires face as the startup grows? Should I compensate early 

hires differently fi^om later hires? 
b . Investor Dilemmas — What types of investors should I target at different stages of 

growth? What challenges will these investors introduce? 
c. Founder-CEO Succession — Why and how are founders replaced as CEOs of the startups 

they founded? How can founders exert more control over the process? What happens to 

the founder and to the startup after he or she is replaced by a hired "professional CEO"? 



Should I found 
now? 

No 



Yes 



Should I be a 
solo founder? 



No 



Remain nonfounder 

Figure 1.1. Sequence of Founding Dilemmas 



Founding -Team 
Dilemmas: 

• Relationships? 

• Roles? 

• Rewards? 



Yes 



Beyond-the-Team 
Dilemmas: 

• Hires? 

• Investors? 

• Succession? 



Figure 1.1 is a high-level summary of these dilemmas and a road-map to which we will return 
regularly. 

These dilemmas do not necessarily occur in the clear sequence implied by the book's structure. 
The founding process is often chaotic and nonlinear,^ with founders improvising rather than following 
a script to build their startups.^ For example, core founders often get an idea — the proverbial light 
bulb goes off — and then face the question of whether to pursue it alone or attract cofounders. This is 
the "idea-then-team" approach to building a startup, exemplified in this book by Tim Westergren (of 
Pandora Radio) and others. But it can also happen that a team decides to work together on a startup, 
then hunts for an idea. This is the "team-then-idea" approach, exemplified by Janet Kraus and Kathy 
Sherbrooke, who met at Stanford Business School, formed a close working relationship through joint 
leadership roles, and decided that they would eventually try to find an idea with which to found a 
company. That startup later became the corporate-concierge company Circles. Given such differences 
in sequence, I've written this book so that people who want to learn about founding dilemmas in 
sequence can read the chapters in order while people interested in specific dilemmas can go straight 
to the chapters that most interest them 

More important than the sequence is what the dilemmas have in common: They are all difficult but 
necessary decision points, each with a number of often-unrecognized options that, in turn, have 
important consequences for the founder(s) and the startup. They each call for carefiil and rational 
decision making — sometimes by a lone founder and sometimes by a founding team. Finally, the 
dilemmas examined in this book all hold in common their tendency to manifest the following three 
major themes. 



Short-term versus Long-term Consequences 



An "easy" short-term decision may introduce long-term problems; a "hard" short-term decision may 
often be best in the long run. Conflict avoidance often leads ft)unders to make easy short-term 
decisions; they succumb to the tenptation to sidestep or postpone acknowledging — not to mention 
resolving — these dilemmas, especially if coming to a decision would require difficult conversations 
about what could go wrong. Let's suppose you have started a company and made your brother the 
chief financial officer, as did one of the founders we will study. What if your brother turns out not to 
be very good at it? The only thing more awkward and awftil than talking about that in advance would 
be dealing with it once it's happening. Founding a startup is akin to a wedding, a declaration of 
mutual devotion. It seems inappropriate and even counterproductive to plan for a breakup, yet in 
entrepreneur ship, failing to make the prenup part of the wedding vows, so to speak, can prove 
disastrous. 

To make matters worse, entrepreneurs usually find it much harder and more costly to undo an early 
founding mistake than to make the right decision to begin with. Sequences of decisions are also often 
path dependent; early decisions tend to close off fiiture options that the founder would otherwise 
have regarded as atttactive. We will examine costly mistakes that should be avoided and, where 
possible, explore ways to undo such mistakes. 

Members of the entrepreneurial community often say that the founding process is "a marathon made 
up of a series of 100-yard dashes." Every 100 yards, the runner/founder faces a key decision with 
long-term consequences. Unless you have a clear picture of the path to the finish line, decisions made 
during any specific dash can throw you off course or put you out of the race altogether. 

Natural Biases: The Perils of Passion, Optimism, and Instinct 

Founders tend to pride themselves on being action-oriented and optimistic — necessary traits, indeed. 
A founder's passion is essential to launching a startup, but it can become deadly at almost every step. 
Likewise, founders' natural biases — toward optimism over realism, toward instinct over systematic 
planning, toward strong attachment to their ideas, their startups, and their employees over 
dispassionate reasoning — often turn on them 

Take optimism. Excessive confidence and optimism in the startup's prospects can lead 
entrepreneurs to involve family and friends both as employees and as investors, imperiling both the 
relationships and the startup. Optimism leads founders to make overly rosy predictions about their 
own chances of success in comparison to their competitors' chances,^ to overestimate their own 
abilities and knowledge,^ to underestimate their initial resource requirements, and to fail to plan for 
foreseeable problems. As a result, founders often fail to attract the resources they will actually need, 
increasing their chances of failure. Too much hubris, confidence, and passion can hinder a founder 
from exploring alternative approaches and making necessary adjustments.^^ In addition, the 
entrepreneurial environment itself can exacerbate a founder's vulnerability to his or her own 
overconfidence. Many entrepreneurs work in highly uncertain and "noisy" environments in which 
feedback is ambiguous and based on uncertain evidence. Such environments render people especially 

susceptible to cognitive biases and errors, including overconfidence,^^ and to following their 
instincts, the one "inpuf that does seem clear. 

As we will find in each chapter, founders need to see past their instincts and their natural 
propensity for wishful thinking to grasp the fiill range of options and consequences. They need to 
expect the best while preparing for the worst and to make decisions strategically rather than 



reactively. Founders may repeatedly find that simply following their instincts will prevent them from 
thinking hard enough about their decisions and the consequences of particular paths of action. 

Divergences among Can, Do, and Should 

Thanks to conflict avoidance, path dependence, natural biases, and ignorance of the long-term 
consequences of their decisions, founders often take actions that diverge considerably from what they 
should do. In each chapter, we will explore the range of available options (what founders can do) 
and how often founders choose each option (what they do). Whenever possible, we will examine how 
decisions affect inportant outcomes, such as the startup's growth, the founding team's stability, and 
the founder's long-term control; this will offer insight into what founders should do. Each chapter 
will close with prescriptive recommendations: how founders can make better decisions. 

First- time founders will likely not know about the dilemmas we will examine. They may not 
understand the fiill range of options available to them, and they may not appreciate their decisions' 
long-term and often cumulative consequences. (They may not even realize they are making an 
important decision!) But even experienced founders can blunder, especially if their experience has 
afforded a vivid but limited picture of the founding dilemmas. 

WEALTH VERSUS CONTROL: A CLOSER LOOK 

Now that we've introduced the key concepts and arguments relating to founders' dilemmas, let's 
probe deeper into the most common and diflRcult dilemma of them all. Founders may misunderstand 
not only the rules and moves of the game, but also their goals in playing it. Many admire founders 
who built and still run their companies, startup founders who became rich and powerfiil CEOs. They 
remember Bill Gates, the geeky teenager who became the fearsome CEO of one of the world's 
biggest and most aggressive companies, not to mention the world's richest person. Yet it is the rare 
founder who achieves this "entrepreneurial ideal" of building a business, running it, and getting rich 
from it. In fact, a dirty little secret of entrepreneurship is that many decisions along the 
entrepreneurial journey — even unquestionably "righf decisions — tend to push that ideal a little 
farther out of reach by forcing the founder to choose between building the company's value and 
maintaining his or her own control over it.* Evan Williams faced these trade-offs with each step he 
took building Blogger and Odeo. The very different decisions he made in each case brought about 
very different outcomes precisely because he had struck different trade-offs. 

Although the desires for wealth and control seem complementary, as entrepreneurial motivations 
they turn out to exist in perpetual tension with one another. This counterintuitive clash results from the 
central challenge entrepreneurs face — the "resource-dependence" challenge. ^-^ Founders need to 
attract outside resources — people, information, and money — in order to pursue opportunities and 
build the greatest value. But acquiring these resources typically requires that founders cede more 
and more control. Cofounders and key employees want equity. Skilled and experienced employees 
want to call some of their own shots. Investors want to protect their investments, typically by means 
of a strong presence on the startup's board and by other terms of the deal structure. 

With a wide range of possible options, how can entrepreneurs choose the best one? The key is to 
know what they want to accomplish with any given decision. Similarly, when founders must choose 
over and over again between profiting from their hard work and keeping control of their own 



creation, the best way for them to know what to do is to know their own motivations — why are they 
there in the first place? An individual's motivations are almost always complex and rarely 
transparent to anyone — including the individual himself or herself Still, research shows that 
founders' two most common motivations are (a) building wealth and (b) driving and controlling the 
growth of their startups. Founders who consistently make decisions that build wealth are more 
likely to achieve what I call a "Rich" outcome (greater financial gains, lesser control), while 
founders who consistently make decisions that enable them to maintain control of the startup are more 
likely to achieve what I call a "King" outcome (greater control, lesser financial gains).* 

Multiple studies have confirmed that these two motivations are the most common. In the Kauffimn 
Foundation's study of 549 founders of American technology startups, 75% of the respondents said that 
building wealth was a very important motivation for becoming an entrepreneur and 64% said the 

same of wanting to own their own businesses. Likewise, the Panel Study of Entrepreneurial 
Dynamics asked 1,214 respondents about their motives for starting a business. The top six 
motivations were control motivations, such as fi-eedom to take one's own approach to work and 
fiilfilling a personal vision, and wealth-building motivations, such as gaining financial security and 

building great wealth. The CareerLeader database, which includes more than 2,000 entrepreneurs 
(and 27,000 executives in total) worldwide, found that for male entrepreneurs in their 20s and 30s, 
the top four motivators all amounted to forms of control or wealth gain."^ Other motivations, such as 
intellectual challenge, altruism, and prestige, can prove important, but across entrepreneurs as a 
whole, wealth and control compose the big two, and are also the two that repeatedly come into 
conflict with each other. ^ 

A founder who knows whether wealth or control is his or her primary motivation will have an 
easier time making decisions and can make consistent decisions that increase the chances of reaching 
the desired outcome — Rich or King. Figure 1.2 summarizes the core wealth- versus-control dilemmas 
we will examine throughout Parts II and III and that we will revisit in the book's concluding chapter. 
(In Chapter 11, we will also explore the potential advantages and risks of making deliberately 
inconsistent decisions and of trying to maximize both wealth and control.) Yet the motivations 
themselves do not necessarily remain fixed. Founders must also watch for changes in their own 
motivations; we will consider the factors that cause such changes and how they should affect a 
founder's subsequent decisions. 

UNSOLVED PUZZLES 

One test of any new framework or model is whether it can explain unanswered questions or unsolved 
"puzzles." This book will tackle puzzles about entrepreneurs, including the puzzle of "the missing 
private- equity premium" and the notion that founders aren't as powerfiil within their companies as we 
commonly assume. 

Most economists assume that entrepreneurs are out to make a lot of money. According to one 
researcher, "That the entrepreneur aims at maximizing his profits is one of the most fiindamental 

assunptions of economic theory." One recent study supports the notion that entrepreneurs want to 
build personal wealth by owning equity in a valuable startup, finding that many founders of high- 
technology startups believe they stand a much greater chance of becoming wealthy by launching a 
startup than by ordinary employment.^^ 



If many founders really do believe this, it appears they are largely wrong. On average, 
entrepreneurs earn no more by founding startups than they would have earned by investing in public 

equity — less, in fact, from a risk-return perspective.^^ One large study of startups suggested that, on 
average, it would not make sense for potential entrepreneurs with a normal amount of risk aversion 

and a moderate amount of savings to become founders of venture-backed startups. Likewise, for the 
self-employed, initial earnings are lower and earnings growth significantly slower than for those 
engaged in paid employment, a finding that has been confirmed in a variety of industries and using a 

variety of earnings metrics. All told, entrepreneurs earned 35% less over a 10-year period than they 
could have earned in a "paid job. "^"^ The authors of these studies therefore wondered why so many 
intelligent people would wish to play this game if there is indeed no "private- equity premium "^^ In 
Chapters 2 through 10, we will find pieces of the answer, while Chapter 11 offers a comprehensive 
answer: Founders face decision after decision in which they must choose between increasing their 
wealth and maintaining control. Founders who choose to remain King should indeed end up less Rich. 

Putential 



Pjrtidpjnlt 
in the Sliirtup 


Decititm Area 


Deiisiims OrietiteJ ti/titirj 
Mjifttaitmtg Control 


Dteisions Oriented totvard 
Maximizing Wealtit 


Cofoundcrs 


Solo vs. team 
Relationships 
Roles 


Remain solo founder (or attract 

weak cofoundersl 
First look to immediate circle for 

"comfortable" cofounders 
Keep stronji control of decision 

making; build hierarchy 


Build founding team; attract best 

cofounders 
Tap strong and weik tKs to find the 

best (and complementary) cofounders 
Give deci&ion-roaking control to 

cofounders with expertise in specific 

areas 

Share equity to attract and/or motivate 

cofounders 




Rewards 


Maintain most or all equits- 
ownership 




Relationships 

Roles 
Rewards 


Hire witliin close personal ncfM'ork 
(friends, family, and others) 
as required 

Keep control of key decisions 

Hire less expensive junior cmplo)-ees 


Aggressively tap broader network 
(unfamiliar candidates) to find the best 
hires 

Delegate decision making to appropriate 
expert 

Hire e.vperienced employees and iiKent 
them with cash and equity 


Investors 


Self -fund vs. take 
outside capital 
Sources of capital 


Self-fund; "bootstrap" 

Friends and family or monc>'-only 
angels; tap alteriutivc sources (e.g., 
cuttumer prepayments ordebc) if 
p<«vsible 


Take outside capital 

Target experienced angels or s^nture 
capitalises 




Terms 

Board of dircctofs 


Resist investor-friendly terms (e.g., 
refuse any supermajoriry rights) 

Avoid building ofticial board; lA'hen 
built, control composinon and 
makeup 


Be open to terms necessary to attract best 
investors (e.g., supermajoriry rights) 

Be open to losing control of board if 
necessary to get best investors and 
directors 


Suixcuors 


Trigger of succession 

Openness to succession 

Desired role after 
succession 


Avoid succession issue until forced 

Resist giving up the CF.O position 

Prefer to kravc than to remain 
"prince* 


Be open to imtiJtuig succession when 

next srjge of startup is outside one's 

own expertise 
Be open to giving up CEO position to 

better CF.O 
Want to remain executive in position 

that matches skilb and preferences 


Other betort 


Preferred rate of ■startup 

g/owth 
Capital mtensir>- 
Corc founder's "capitals* 


Gradual to moslcr.ilc 

Low capital intensity 
Wcll equipped to launch and build 
startup without much help 


Fast to explosive 

High capital intensity 
Important gaps that should be filled 
by mvolving others 


Most likely 
outcome 




Maintam control; build less value 


Build financial value; imperil control 



Figure 1.2. Wealth-versus-Control Dilemmas 



Turning to the question of founder power, academic studies have assumed that the founder of a 
startup — the person who conceived the idea and assembled the people and resources — looms as a 

powerful person within that startup.* How could the "revered founder" not be?^^ Yet once again, the 
evidence described here belies a fundamental assunption. My quantitative research found that a high 

percentage of founder-CEOs are replaced as CEO, most often against their wills. The public thinks 
of business eminences such as Bill Gates, Richard Branson, Anita Roddick, and Michael Dell — 
wealthy and powerful CEOs of the companies they created — but these are the very rare exceptions. 
Far more common is the founder-CEO who is replaced before his or her startup gets anywhere near 
the public markets. How can we explain this? And why is it that founders receive, on average, 
significantly /owr compensation than nonfounders?^^ Again, these seemingly puzzling facts will 
make sense by the time we revisit them in Chapter 1 1 . 

In much of the academic literature, wealth and power go hand in hand. In C. Wright Mills's classic 
analysis of the "power elite," for instance, the corporate power and economic wealth of top 
executives reinforce each other, with corporate positions serving as the source of wealth while 
"money provides power."^^ However, my research has found that wealth and power are decoupled 
for entrepreneurs and, indeed, in active conflict. As a result, few founders of high-potential startups 
can achieve both wealth and power; most choose between one or the other and often end up with 
neither. Even a seemingly reasonable strategy for achieving high levels of both wealth and power, 
rather than maximizing one over the other, is actually more likely to put them both out of reach. We 
will see why this frustrating condition occurs. 

10,000 FOUNDERS STRONG 

To paint a rich but rigorous picture of the most important decisions made by founders and the 
outcomes that founders experience. The Founder 's Dilemmas marries deep case studies with 
analyses of a unique large-scale database.* Given the lack of comprehensive public data sources 
about startups, I collected my own data from across the United States by performing annual surveys of 
private high-potential startups. The survey included questions about each startup's founders, 
nonfounding executives, compensation and equity holdings, financing history, board of directors, and 
other dimensions of organizational life. I conducted these surveys for 10 consecutive years, from 
2000 through 2009, creating a unique dataset that includes 9,900 founders — and more than 19,000 
executives in total — from 3,607 startups. These data are more representative of high-potential 
startups in the United States than any other dataset of comparable size and richness. The decade 
covered in the surveys spanned all stages of the business cycle, from the heights of exuberance in 
Internet startups to the depths of despair and pessimism, and back. As such, it forms the quantitative 
backbone for every chapter of this book. Appendix A provides detailed breakdowns of the startups 
that participated in the surveys as well as the survey questions posed to entrepreneurs. 

The dataset and case studies both focus on the two most central industries for high-potential 
startups, technology and life sciences.* Together, these industries dominate every measure of young 
startup employment and funding. Of the initial public offerings (IPOs) during the decade (2000- 
2009), 48% came from those two industries, and no other industry accounted for more than 12%.^^ 
Furthermore, of the angel capital invested during the decade, 74% went to those two industries,^ ^ as 
did 71%) of venture capital. Many more low-tech "small businesses" (themselves the subject of a 
growing literature) start up each year than high-potential startups, but most of those small businesses 



lack employees and do not intend to grow or innovate. ^ This leaves high-potential startups as the 
core of the economic growth associated with entrepreneur ship. 

My focus on technology and life sciences startups in the United States allows us to examine the 
tensions between wealth and control, particularly potent in these startups because of their intense 
need for outside resources and because of the terms under which entrepreneurs typically acquire 
those resources. Including large numbers of both technology and life sciences startups enables us to 
compare the data and dynamics from those very different industries and assess which patterns are 

industry-specific.^ Where we find patterns common to both industries, we can feel more confident that 
they generalize to startups outside these two industries. On the other hand, the degree to which the 
patterns described here apply outside the United States is an open question that deserves empirical 
attention. A lack of financing options in many other countries may preclude founders from considering 
some of the investor-related decisions described in Chapter 9, while a lack of experienced executives 
might preclude some of the hiring dilemmas described in Chapter 8. In other industries, such as those 
that are not capital-intensive or those that can attract customer revenues before they incur expenses 
(such as those in which subscription-based business models proliferate), founders may not have to 
consider outside financing options because they can fiind their startups themselves. The applicability 
to those industries of the dilemmas described here is likewise an open question, one taken up in detail 
in Chapter 1 1 . 

INTRODUCTION TO THE PRINCIPAL CASE STUDIES 

Of the more than three dozen case studies of founders of high-potential startups we'll examine, I 
chose to highlight seven across multiple chapters, both for the insights gained from the dilemmas 
faced by the founders and for the window these case studies collectively open on a wide range of 
circumstances, dilemmas, decisions, trade-offs, and outcomes. Figure 1.3 summarizes these seven 
startups in terms of their founders' backgrounds, their founding teams, and their hiring and investor 
decisions.* Let's briefiy meet these six entrepreneurs (in addition to Evan Williams, whom we've 
already met) and introduce the dilemmas they faced in starting and running their businesses. 

Pandora Radio's Tim Westergren is as unlikely a founder as you could imagine. He studied 
piano and political science at Stanford, then worked as a nanny, ran an admissions office at Stanford, 
and started a rock band, touring with it for nearly a decade before deciding to move on to freelance 
composing. During his composing years, Tim had the idea to create a music database (which he 
called the "music genome projecf ') that categorized music based on a long list of attributes and that 
suggested to users new artists and songs that fit their tastes. In 1999, Tim met Jon Kraft, a Silicon 
Valley entrepreneur. The two decided to pursue Tim's idea and formed Pandora (originally called 
Savage Beast), adding Will Glaser, a talented software engineer, as the third cofounder and CTO. 
The team split the equity equally, adopted a clear division of labor that matched their very different 
areas of expertise, and gave each founder autonomy to make hiring and business decisions within his 
own ftmction. However, as the conpany burned through its dwindling cash and had to defer founder 
and employee salaries, these early decisions about roles and team structure heightened tensions 
within the team and exacerbated personal problems, leading Jon to leave and causing problems for 
the remaining founders. 



Prior 

Relationship 



Startup 


First -Time 


Solo 


with 






(Core Founder) 


Founder^ 


Founder^ 


L,ofotinders 


Hiring 


Investors 


Pandora Radio 


Yes 


No 


Acquaintances 


Friends 


Friends 


(Tim Wcstcrgrcn) 












Masergy 


Yes 


Yes 


N/A 


Young, 


Top 


(Barry Nails) 








replace 
as 

startup 
scales 


partners 
from top 
VC firms 


Sinartix 


Yes 


No 


Classmates 


Pre- hiring 


Pre- 


(Mvek Khuller) 






(and one prior 
coworker) 




funding 


Sirtercity 


Yes 


Yes 


"Couplc- 


Young, 


Angel 


(Genevieve 






preneur" 


replace 


investors 


Tliicrs) 








as startup 
scales 




Ockham 


Yes 


No 


Prior 


N/A 


Deciding 


Technologies 






coworkers 




on angel 


(Jim Triandiflou) 










vs. \'C 


Blogger/Odeo 


Serial 


No 


Various 


\'arious 


Various 


(Evan Williams) 


founder 










Feed Burner 


Serial 


No 


Prior 


\'arious 


Various 


(Dick Costolo) 


founding 
team 




coworkers 







Figure 1.3. Principal Case Studies 

Masergy's Barry Nails also had an intriguing background. Some in the entrepreneurship world 
might call him a late bloomer — something he acknowledges but regrets. Barry joined GTE, a large 
telecommunications company, right out of school and over the next 25 years steadily climbed the 
ranks, gaining deep sales experience, profit- and- loss responsibility, and other executive skills. When 
GTE announced merger plans, he left to work for a couple of young companies. He finally took the 
leap himself and founded a telecom service provider called Masergy. A self-described "solo guy," 
Barry decided to be the sole founder and CEO of Masergy and used a six-month severance from his 
last enployer to fimd his startup until finally raising millions of dollars from VCs. Although his 
startup benefitted from his wealth of prior work experience, it was only after taking the founder's 
leap that Barry learned key lessons about hiring a team for a fast-growing startup, managing a board 
of directors, and negotiating the tensions between sales and operations. 

Vivek Khuller of Smartix, our next founder, struggled with the other side of the problem faced by 
Barry: Would a young MBA student be able to revolutionize an industry about which he knew next to 
nothing? While Vivek was working as a summer intern at an investment bank, he conceived the idea 
of using electronic-ticketing technology to allow sports and entertainment venues to issue and process 
their own tickets, thus eliminating intermediary ticketing agencies such as Ticketmaster. Mvek 
decided to advertise for cofounders within his MBA program, choosing two bright classmates with 



backgrounds similar to his own. Through his business school contacts, Mvek was able to pitch his 
idea to several high-profile venues, which were interested in partnering with Smartix, and to several 
top-tier VC firms, one of which issued a term sheet. However, his early founding decisions soon 
came back to haunt him. 

For Genevieve Thiers, founder of Sittercity, the biggest dilemma was also highly personal: Her 
most inportant coworker was also her fiance, Dan. What would happen to the business if she and Dan 
broke up? And what would happen to her relationship with Dan if their business relationship soured? 
The epiphany behind Sittercity had come to Genevieve as she was helping a pregnant woman post 
fliers for a babysitter. Reaching back to her own e?q)eriences as a professional babysitter, Genevieve 
realized that a website allowing parents to quickly access babysitters could be a winning idea. After 
graduation, she worked as a technical writer at IBM and pursued her avocation as an opera singer, 
developing Sittercity in her spare time. When IBM shut down her division, Genevieve focused on 
Sittercity full-time. As the startup grew, she enlisted Dan as a technical advisor, and the two worked 
closely to build the business. Dan eventually became COO, but Genevieve was nervous about the 
arrangement. She needed to find creative and essential ways to erect "firewalls" to protect herself 
from the inherent dangers of her early choice of a coworker. 

By contrast, Jim Triandiflou of Ockham Technologies e?q)erienced a dilemma that most 
entrepreneurs would envy: choosing from a number of interested investors. Jim suspected that the 
decision would deeply affect the conpany's future and his own role in it. Should he choose the angel 
investor who came with few strings attached or the VC firms that expected more control but would 
provide needed resources? Jim was the son of schoolteachers; his father was risk-averse, having 
worked in the same classroom for 33 years, and Jim saw himself in the same mold. With a degree in 
marketing and an MBA, Jim had been working at a consulting firm and not even thinking of taking the 
plunge as an entrepreneur. But after talking to his colleague Ken about Ken's idea for a sales- 
management software startup, Jim decided to give it a try. Jim and Ken recruited Mike, who had 
worked for Jim at the consulting firm, as the third cofounder of Ockham Technologies. The arrival of 
his first child convinced Ken not to quit his job to join Ockham after all, leaving Jim and Mike to 
proceed on their own. In short order, they landed IBM as a customer and used an outsourcing 
conpany to develop their software. As CEO, Jim was successful at pitching the concept to investors, 
but each investor's offer would hold significant implications for Ockham' s growth and development. 

The book's final primary case study focuses on Dick Costolo, founder of FeedBumer and three 
prior startups and, subsequently, CEO of Twitter. After collaborating on several startups, Dick 
and his "serial founding team" seemed to have developed a winning formula for low-tension, high- 
value cofounding. Before embarking on his business career, Dick had been a Chicago stand-up 
comedian. After a few years working as an engineer at Andersen Consulting, he became finstrated by 
Andersen's lack of interest in pursuing Internet technologies and left with a colleague to start a 
website consulting company. In their early startups, the team ran into problems with poorly chosen 
hires, with investors, and with each other. But the founding team stuck together, in the process losing 
one cofounder while adding three others. In their fourth attenpt, FeedBumer, they seemed to hit on all 
cylinders as they applied lessons from their prior startups. They were within striking distance of the 
Promised Land when they learned that even a smoothly functioning team can fall into disagreement as 
it nears the finish line. After solving those dilemmas, Dick went on to become part of one of the 
highest-profile founder-CEO succession events ever, taking the helm at Twitter. 

These founders appear throughout the book. Other founders, such as a husband-and-wife team who 



founded a life sciences startup, play important roles in individual chapters. As a group, these 
founders are extremely varied, with very different personalities, styles, backgrounds, and abilities. 
Yet, as we'll see, the dilemmas they face are surprisingly similar. All must decide when to make the 
leap into founderhood, whether and how to cofound with others, when and how to hire nonfounding 
employees, and whether to self-fund or raise outside capital. At each fork in the road, the wrong 
decision can send the startup over a cliff or bring about the founder's replacement as leader of the 
expedition. Less dramatically, each outcome also heightens the startup's chances of success or 
failure. 

Founders embarking on their entrepreneurial journey have had no roadmap to follow, no way to 
anticipate the forks, potholes, and ditches they may encounter. By arming entrepreneurs with a 
founders' roadmap, I want to point out common pitfalls, improve the stability and effectiveness of 
their teams, and help them reach their desired destinations. Our three dozen case studies will map the 
terrain, and our data on 10,000 founders will show us how often each decision is made and with what 
implications. Founders who gain insights into their core motivations for embarking on this perilous 
startup journey should also gain understanding of the trade-offs they will have to make each step of 
the way and which choices will help them reach their destination. 

The first fork in the road is the pre-founding decision about whether and when to embark on the 
journey. As I will argue, when three factors — career factors, market factors, and personal factors — 
are all favorable, the decision is easy. However, potential founders who encounter one or more 
unfavorable factors grapple with real dilemmas about when to found, potentially imperiling the 
startup, their family situations, or their ability to embark on an entrepreneurial journey in the first 
place. These three factors, and the ways in which they should affect the decision to found a conpany, 
are the focus of Chapter 2. 



* In a classic two-page discourse, Stinchcombe (1965) argued that the liability of newness was due to three internal factors — the 
needs for the team to develop working relationships, to find their new roles, and to split financial rewards among themselves — and one 
external factor — the lack of relationships with potential suppliers, customers, and other external parties. The last factor has received vast 
amounts of attention, the first three very little. 

* This is not to say that choice of strategy, business model, and industry segment are not important. We will indeed consider the 
effects of those choices when appropriate. 

* Garland et aL (1984:354) state, "[AJlthough there is overlap between entrepreneurial firms and small business firms, they are 
different entities." Aldrich et al. (2006) differentiate between the small businesses on which they focus and "high technology" startups, 
such as those studied in the Stanford Project on Emerging Companies (SPEC). Yet, many of SPEC'S startups were outside traditional 
"technology" industries, as is the case with the life sciences startups in my dataset, necessitating the use of a broader term for such 
organizations. (Thus, I use "high-potential startups.") Schumpeter (1934) provides us with another way to differentiate between small 
businesses and "entrepreneurial" businesses. He characterizes new entrepreneurial businesses as being geared toward introducing new 
goods, introducing new methods of production, opening new markets, and/or opening new supply sources. See Carland et aL (1984) for 
further discussion of these characteristics and differences. 

* Stevenson's definition of entrepreneurship has also been applied to people who did not start their organizations but are "acting 
entrepreneurially" in building those organizations. In contrast, "founders" are the people who start their organizations. 

* Parts of this trade-off have parallels in at least two other academic literatures. First, the finance literature has looked at the "private 
benefits of control," using such measures as the difference in the prices of securities that are identical except for their voting rights, and 
found that people are willing to give up economic gains to gain control or voting rights (e.g., Grossman et aL, 1988; Lease et aL, 1983). 
Second, the negotiations literature (e.g., Walton et aL, 1965) has examined some elements of the theoretical tension between value 
claiming and value creation. 

* This is one of the areas in which owner-operated small businesses differ significantly from high-potential startups, which require 
significant resources to reach their full potentiaL 



^ The very extensive CareerLeader survey, which was developed by psychologists and psychometricians Dr Tim Butler and Dr. 
James Waldroop, also evaluates a person's core interests, the kind of organizational culture he or she most enjoys and succeeds in, his or 
her strengths and weaknesses, and the characteristics that may limit his or her success. Chapter 2 lists the 13 potential motivators and 
details the breakdown of motivators by entrepreneur/non-entrepreneur, gender, and age group. 

i Founders often talk broadly about starting a business because they want to "have an impact." Yet what they mean by "impacf can 
vary widely. Control-motivated founders might dream of bringing to market a product that fully realizes their uncompromised vision, even 
if it means that fewer people will buy it. Wealth-motivated founders might dream of getting a product into as many hands as possible, 
even if the founder wound up having little influence over the product's characteristics and features. Diverging ideas about impact can — 
and should — lead founders to make very different decisions and take very different actions. 

* For instance, in his seminal study of the dimensions of executive power, Finkelstein (1992:509) posits that founders not only are 
powerful figures within the startup team itself, but "may gain power through their often long-term interaction with the board, as they 
translate their unique positions to implicit control over board members." The power of the founder is assumed to be so strong that it even 
enhances the power of the founder's relatives who are involved in the startup. 

* For a discussion of when "hybrid" methods (those that integrate qualitative and quantitative methods) are appropriate for domains in 
which our knowledge of a phenomenon is still developing, see Edmondson et aL (2007). 

* I will usually focus on the commonalities across these two industries, but at times I will also use the technology startups as a baseline 
against which to observe differences between them and life-sciences startups (e.g., in the preponderance of solo founders and in the 
amounts of capital they raise from investors). 

^ But even for patterns that are common to both information technology startups and life sciences startups, we cannot conclude that 
they are universal without fUrther cross-industry analysis. 

* Full-length Harvard Business School case studies are available from HBS Publishing for each of the principal case studies below and 
for most of the other cases discussed in this book. Appendix B provides fuU citations for those cases. 



CHAPTER TWO 



CAREER DILEMMAS 



Potential founders face a variety of pitfalls when deciding whether, when, and how to found a 
startup. They may dive into launching a startup without thinking about whether they have the skills and 
motivations necessary to succeed; they may focus on business issues without considering their 
personal situations and how those situations may undermine their success (or be undermined by it); or 
they may fall in love with the idea of founderhood and enthusiastically latch onto an exciting idea 
without dispassionately evaluating it. Founders who dive in too early may doom themselves to 
destructive failure. Those who wait too long may find themselves handcuffed by a high-priced 
lifestyle and a steady paycheck, then look back at the end of their careers and regret having never 
made the leap. 

To help us understand the wide range of career options, let's meet two potential founders who 
faced very different circumstances and made very different decisions. 

Hunphrey Chen was a potential early-career founder. Although he came fi"om a traditional 
Taiwanese family that strongly encouraged him to pursue a career in medicine, he was drawn instead 
to startups and cutting-edge technology; he loved the "excitement and uncertainty" of rapid-growth 
industries and businesses. While still in the second year of an MBA program, Humphrey and a 
classmate developed a new technology to enable a listener to identify and purchase music that was 
playing on the radio. "The idea just lit up in my head," said Humphrey. ^ By dialing a phone number 
and inputting the station number while the song was still playing, listeners could immediately order 
the song. Humphrey was so excited by his idea that he filed a patent, developed a basic technology 
demo, and incorporated a company he called ConneXus to monetize the idea. 

Meanwhile, a top consulting firm had offered Humphrey a lucrative job working on innovative 
media and communications projects in Manhattan. The job fit nicely with his interests and 
background. Adding to his dilemma, Humphrey was getting married in a few months to Cecilia, who 
was working toward her pharmacy degree and not earning a salary yet. Cecilia explained, "I'm not as 
bold as Humphrey ... I am more of a stable person. The consulting job would offer security and I 
would definitely feel comfortable with that."^ Humphrey, however, did not want to turn away fi-om 
his startup idea and began negotiations with investors, torn about whether to pursue launching 
ConneXus full-time or to take the consulting job. 

In contrast, Barry Nails waited much longer before considering whether to make the leap. Barry 
came fi*om a long line of small-business owners in Texas and had been, as he said, "surrounded by 
entrepreneurs and small businesses my entire life." But instead of following in his father's and 
grandfather's footsteps, Barry earned a two-year technical degree and then went to work for GTE, 
one of the largest companies in Texas. Over the next 25 years, Barry climbed GTE's corporate 
ladder, tackling increasingly more responsible positions. But when GTE announced in 1999 that it 
would be merging with Bell Atlantic, Barry's worries about what the merger might mean for him 
spurred him to think more deeply about starting his own business. He was married with two school- 



aged children, but his wife encouraged him to make the leap and was even willing to move to make 
the transition easier. After first working for two startups and gaining knowledge of what that kind of 
life was like, he found his wife still supportive of his becoming a founder himself and began looking 
around for an idea with which to start a company of his own. 

Do most founders start at a young age, like Humphrey, or are they later bloomers like Barry? 
Neither. Stage of life does not seem to be a strong factor in starting one's own business. In my dataset 
of thousands of first-time founders of technology and life sciences startups, there was wide variation 
in the stage of life during which founders decided to make the leap; there was no "sweet spot" of 
founding ages.-^ On average, founders had worked for 14.0 years before making the leap, but with a 
large standard of deviation of 9.8 years. Technology founders averaged 13.1 years of prior work 
experience and life sciences founders averaged 15.9 years. A full 35% of founders had worked 20 
years or more before founding, including 47% of life sciences founders, but a distinct subset founded 
with only 0 to 4 years of work experience.^ Humphrey would fall at the very low end of this age 
distribution and Barry at the high end.^ But while they were quite different regarding the ages at 
which they had the urge to found a company, the factors they considered when deciding whether and 
when to act on that urge were remarkably similar. 

Before following their passions, potential founders should step back to answer the pre-founding 
questions we will tackle in this chapter: 

1 . Should I become an entrepreneur? 

2. If so, when should I make the leap into founderhood — early in my career or after I 
accumulate more career experiences? 

3. How can I dispassionately evaluate my idea? 

These questions are particularly complex because they mix personal issues ("Will my spouse be 
supportive when things get tough? Is my family situation conducive to founding a startup?"), career 
issues ("Have I accumulated skills and experiences that prepare me to found a company?"), and 
market issues ("Is my idea good enough, or is my passion for it misleading me?"). Each founder's 
answers to these questions will have important repercussions for all of the subsequent dilemmas he or 
she will face. Early-career founders, for example, will need to consider adding cofounders and 
experienced hires to make up for their own missing skills. Late-career founders who accumulate 
relevant pre-founding experiences, on the other hand, may be in a better position to consider solo 
founding and hiring less-experienced people. 

SHOULD I FOUND? 

At almost any age, potential founders may be struck by a high level of passion, whether for a business 
idea itself or for the idea of becoming a founder or for the chance to change an industry. That passion 

is often a critical source of entrepreneurial motivation,^ and it is common to all the founders 
discussed in this chapter. That said, the founders examined here differ in almost every other way: age, 
experience, childhood influences, personality, marital and family status, economic status, and where 
their ideas came from Are there patterns in the way people decide to become founders? 

Despite the popular conception of entrepreneurs as risk takers, risk aversion does not seem to 
affect the decision to become an entrepreneur. Academic studies testing for a significant difference in 



risk aversion between entrepreneurs and non- entrepreneur managers have produced conflicting 

results or suffered from methodological problems. However, as detailed below, studies have 
suggested that two other factors may separate entrepreneurs from non- entrepreneurs. First, early 
influences can have a powerM effect on people who are considering becoming founders, pushing 
some toward founderhood and pushing others away from it. Second, these influences can be 
reinforced by people's natural motivations and the rewards they value. 

Early Influences 

A founder is greatly influenced by the family and culture in which he or she grew up and, in some 
cases, by specific role models. Dr. Tim Butler, who has studied entrepreneurial career issues for 
more than two decades, says that the most powerM, though often unnoticed, influences may come 
from the early messages sent by the words and actions of older relatives or by the culture in which a 
person grew up: "We receive very powerfiil messages about what's important, what success is, what 
failure is, what counts for achievement and what doesn't. In many cultures, that includes messages 
about the value of owning your own business versus being an employee, which shapes our career 
decisions." Those messages can push strongly against becoming an entrepreneur. Humphrey Chen, for 
example, grew up in a family that did not value entrepreneurs. "My parents have always been against 
my career choices. My dad's a doctor and both of my brothers are doctors. In Taiwan, you go into 
business only if you have failed at becoming a doctor [or lawyer or another professional] ... it was a 

different class. ... I was the black sheep in my family."^ This negative message weighed heavily in 
Humphrey's thinking about whether to become an entrepreneur or to follow a more traditional path. 
In contrast, people whose parents, close friends, or neighbors had been self-enployed are more 

likely than others to become entrepreneurs.^ Barry Nails learned to appreciate business ownership 
from the entrepreneurs in his family. "My father owned gas stations and a gun company. My 
grandfather sold antiques, bought and sold land, and owned a wrecking yard. My uncle owned a 
bulldozer company. So I was surrounded by entrepreneurs my entire life," commented Barry. One of 
his earliest memories was attending a flea market with his grandfather: "I was barely high enough to 
see over a table when my grandfather took me to a First Monday Trades day. He put six items on the 
table: a lantern, a cook stove, an antique gun, and other stuff. He said, 'Ask this price. If the buyer 
asks you to take a lower price, then you can take this much.' He'd go off making trades and I would 
stand there making deals. . . . Having grown up in this environment, I saw the whole makeup of how 
you get an idea, how you move forward with it, how you get customers to care about what you're 
doing, how you treat customers." Even after working for a large company for more than two decades, 
Barry's long-ago experiences at the flea market still exerted a powerftil influence on his inclination to 
become an entrepreneur at some point during his career. 

Although many founders lack such sfrong early influences, people who reflect on their early family 
and cultural influences often gain insight into why they are attracted to or shy away from starting a 
new company. When such influences are combined with analyses of the person's core motivations, a 
more complete picture emerges of whether or not the person should become a founder sometime in his 
or her career. 

Founder Motivations: The Centrality of Wealth and Control Motivations 



Chapter 1 described the CareerLeader data that capture the most common motivations for thousands 
of people worldwide. For each survey participant, CareerLeader analysis rank-orders 13 possible 
motivations, or "work reward values": affiliation, altruism, autonomy, financial gain, intellectual 
challenge, lifestyle, managing people, positioning, power and influence, prestige, recognition, 
security, and variety. 

Working with Dr. Tim Butler, the cocreator of CareerLeader, I separated the 27,000 surveys in the 
CareerLeader database by gender, age cohort (people in their 20s, 30s, or 40s and older), and 
entrepreneurial status (entrepreneur or non-entrepreneur).^^ The differences in motivation between 
entrepreneurs and non- entrepreneurs were striking, but there were fewer differences between the 
genders (we will focus on gender differences later in this book) and even fewer differences among 
age cohorts (especially for entrepreneurs, whose motivations were more stable over the decades). 

The top four motivations for males in their 20s are shown below. For the entrepreneurs in this 
category, the list is dominated by control-related items (power and influence, autonomy, and 
managing people) and financial gain.* Interestingly, with the exception of financial gain, the top- 
ranked motivations for entrepreneurs were much lower ranked for non-entrepreneurs and vice versa. 

The founders we will study span this spectrum of entrepreneurial motivations. For instance, at 
Blogger, Evan Williams certainly felt that having power and autonomy was important, even justifying 
his refusal of a multi-million-dollar acquisition offer. "After four years of pouring my heart into 
Blogger, I saw a lot of risk in giving up control," he explained. In contrast, by the time founder-CEO 
Dick Costolo got acquisition offers for FeedBurner, he was motivated by money: "I'm not an engineer 
any more. To me, best price wins." Frank Addante, too, explained that his motivation for launching 
StrongMail was financial: "I thought about it fi-om the perspective of, 'What's going to keep me 
motivated to stick with it and run this for a period of time?' A big enough equity piece will keep me 
motivated to stay." Each of these founders aspired to build a high-impact startup, but "impacf meant 
very different things to each; to the financially motivated, it tended to mean a large gain in wealth, but 
to the control motivated, it tended to mean that the startup would bring to the world the product or 
service they envisioned. 

Table 2.1 

Top Four Motivations for Males in Their 20s 





Miilc Entrepreneurs in 20s 

(RiVik for Miile Non-entreprcnenrs) 


Mijle Non-entrepreneurs in 20s 
(Riink for Miile Entrepreneurs) 


#1. 


Power &: influence (UIO) 


Security (tie) (U13} 


#2. 


Autonomy (#/ 


Prestige (tic) ^#6/ 


#3. 


Mannging people (#9^ 


Financial gain (U4) 


#4. 


Financial gain (U3) 


Affiliation (UU) 



The top four motivations for females in their 20s are shown below. Once again, with one exception 
(altruism), the top-ranked motivations for entrepreneurs were much lower ranked for non- 
entrepreneurs and vice versa. (Mirroring results for the males, control motivations also dominate the 
list for female entrepreneurs, but financial gain ranks much lower than it does for their male 
counterparts.) Genevieve Theirs, who founded the online babysitting company Sittercity while in her 
20s, noted that autonomy and influence were among her motivations: "I felt very disassociated 
[working for IBM], like I was part of a big machine in which I had no idea where I was going. I 



realized that I needed to understand the big picture to be motivated. . . . [At Sittercity], I was on a 
mission to do something big and beautiful." 

Table 2.2 

Top Four Motivations for Females in Their 20s 





Fe/thilc Entrepreneurs in 20s 

(Rank for Female Non-entrepreneurs) 


Female Non-entrepreneiirs in 20s 
(Rank for Female Entrepreneurs) 


#1. 


Autonomy (UI2) 


Recognition (Ul 0) 


#2. 


Power &: influence (it13) 


Affiliation (^2) 


#3. 


Managing people (#/0j 


Security (m^) 


#4. 


Alrruism (#5/ 


Lifestyle (#//) 



The data show that these motivations are relatively stable throughout life, with most of the top- 
ranked motivations for people in their 20s persisting into their 30s and 40s and beyond. The table 
below shows, for each category (in alphabetical order), which of the top-four motivations from the 
20s age cohort remained top-four motivations in the 30s age cohort and which new motivations 
emerged. For female entrepreneurs, variety becomes a top-four motivation once they reach their 30s, 
but entrepreneurs of both genders experienced no other changes in their top-four motivations. In 
contrast, the motivations for non- entrepreneurs seem to be less stable, with half of the top-four 
motivations changing for each gender. 

As shown below, the motivations for male entrepreneurs change noticeably as they enter their 40s 
and beyond. Two of the motivations (autonomy and power & influence) persist throughout their 20s, 
30s, and 40s, but in their 40s, male entrepreneurs also become motivated by altruism and variety. 
(Male entrepreneurs lag female entrepreneurs by a decade when it comes to prizing variety and by 
two decades when it comes to altruism) For non- entrepreneurs, once again, two top-four motivations 
carry over from the prior decade and two new top-four motivations emerge. The motivations for all 
groups become more similar in their 40s, with autonomy and altruism appearing in all four top-four 
lists and variety appearing in three. 

When Major League Baseball pitcher Curt Schilling began to contemplate life after baseball, he 
named altruism as his top motivation for starting a company. Schilling's philanthropic work on behalf 
of ALS ("Lou Gehrig's disease") had motivated him to "philanthropically change the world." One of 
Schilling's model philanthropists was Bill Gates, who had made billions of dollars in the business 
world and then used much of that money to found a high-impact philanthropic foundation. As a result. 
Schilling decided to found 38 Studios, a massively multiplayer online gaming (MMOG) startup, with 
the hopes that gains from it could be used for philanthropic causes. 



Table 2.3 

Top Four Motivations for the 30s Age Cohort 



?0s- Apr 






Willi'* fsJiiil- 




Cohort 


Entrepreneurs 


Entrepreneurs 


cntrcfyrcneurs 


cntrcfyrcttctirs 


Top 4 


Autonomy 


Autonomy 


Prestige 


Recognition 


inotivi rion^ 

11 IWl 1 V II v^l 19 


r iH'i Hi"!'! 1 <>'iirt 
1 iiif.iiivii.li v^*'* ' 


A IrniKiTi 
A 1.11 1 1113111 




^p<^ii nf V 


VII 1 I 1 WI V / > V 1 


l*l«lll«ICfll 1 


1 v7 V 1 V V 










m f 1 1 Ir'Ili'^f^ 

1 II 11 \M\. 1 Id. 








Power & 










iiirtucncc 








New tc^n 4 




Variety 


Posirioniii!? 


Altruism 


motivations 






Recognition 


Variety 



Table 2.4 

Top Four Motivations for the 40s-plus Age Cohort 



40s- fy Ins 




Female 


Male Non- 


Pen tale Non- 


Age Cohort 


Entrepreneurs 


Entrepreneurs 


entrepreneurs 


entrepreneurs 


Top 4 


Autonomy 


Autonomy 


Recognition 


Altruism 


motivations 


Power Sc 


Altruism 


Securitv 


Variety 


carried over 


influence 


Variety 






from the 30s 










New top 4 


Altruism 


Intellectual 


Altruism 


Affiliation 


motivations 


X'ariety 


challenge 


Autonomy 


Autonomy 



The picture that emerges is that entrepreneurs and non- entrepreneurs are motivated by very 
different rewards, especially early in their careers. Potential founders whose highest-ranked 
motivations overlap with those of either entrepreneurs or non-entrepreneurs can get a good idea of 
whether they should consider founding a company at some point in their careers. For the people 
whose motivations are similar to those of entrepreneurs, the next question becomes: When in my 
career should I found? 

WHEN SHOULD I FOUND? 

For those who are entrepreneurially inclined, it is extremely tempting to become a founder as soon as 
possible. However, given the challenges of trying to build an organization from scratch, would-be 
founders should look before they leap. In particular, they should look at whether their career 
experiences have prepared them for the leap into founderhood. However, there is another side to this 
coin. The chance to become better prepared often leads potential founders to wait before founding, 
but waiting until late in one's career can introduce its own perils. Working for other organizations 
can, in itself, make it harder to eventually found one's own startup. Figure 2.1 summarizes the factors 
that pull in each direction. Below, we will delve into each of these competing factors. 

Reasons to Wait 

People who wait to become founders can use their pre-founding career experiences to prepare 



themselves for the extreme challenges of founding. A well-planned pre-founding career can arm a 
potential founder with the human capital, social capital, and financial capital appropriate to a startup. 



Building Human Capital 

As potential founders progress through their careers, they accumulate more and more of the human 
capital needed to found a startup. In this context, human capital refers to the skills, knowledge, and 
expertise needed to launch and build the startup. It includes both general human capital that is widely 
applicable (e.g., leadership ability or the ability to speak and write clearly) and specific human 
capital that is tied to a specific organization or context (e.g., knowing how a particular product is 
manufactured). Looked at a different way, human capital includes formal human capital acquired 
through schooling (e.g., getting a degree in biomedical engineering or computer science) and tacit 
human capital acquired through work and life experiences (e.g., knowing how to negotiate with 
equipment sales reps).^^ 



Figure 2. 1. Factors Affecting the "When to Found" Decision 

As founders accumulate experiences, they also develop mental models (or schemata) — ways of 
categorizing and making sense of information.^^ A doctor and a marketing expert — each of whom 
wants to found a medical software company — will each look at the same information and make sense 
of it in different ways, considering different details to be more important and to have different 
implications for the planned startup, because each brings a different mental model to bear. Founders' 
approaches to building their startups are powerfiilly shaped by the mental models they bring to those 

startups. Accumulating more experience is far less valuable if that experience does not shape the 
mental model in relevant ways (or, worse, if it shapes the mental model in counterproductive ways). 

Work and School 

Before attending business school, Humphrey Chen had spent about a year each at a wide variety of 
employers — Price Waterhouse, Morgan Stanley, a marketing consultancy, and an Internet-based music 
retailer — gaining some breadth of business knowledge. A broad range of work and educational 

experiences is indeed associated with a significantly higher willingness to become self-employed.^^ 
However, Humphrey's breadth came at the expense of specific deep experience or skills that could 
enable him to fully understand the scope or the value proposition of his bright idea for ConneXus. In 
contrast, Mvek Khuller, another young founder, had worked as an engineer for Bell Atlantic for five 
years before enrolling in business school. He understood much more deeply how a world-class 
organization operated and had a much deeper view of the technologies relevant to his startup, which 
enabled him to develop a technical prototype much more quickly. At the same time, Vivek's 



Found early in career... <r 



^ Wait to found until... 



•...before golden handcuffs get too strong 
•...before family handcuffs get too strong 
•...before becoming too specialized 
•...before becoming too reliant on employer resources 



*... build more human capital 
•...build more social capital 
•...build more financial capital 



background did not include experience in the entertainment venues that were his intended market. As 
we will see below, the lack of one or more "capitals" can create a hole that must be filled when 
building a successful startup. 

While Humphrey lacked an extended career in a corporate setting, he did invest time and financial 
resources in schooling. He had expended considerable effort and had achieved success pursuing a 
technical degree and an MBA at top-rated universities. Attending a targeted program, such as a 
focused business-fimction program, or earning an industry-specific credential can be a shortcut to the 
kind of human capital that others, such as Barry Nails, amass through years on the job. More years of 
schooling is indeed linked to a greater likelihood of becoming a founder; after surveying a wide range 
of evidence. Professor Scott Shane concludes, "Including professional school, getting more education 
increases the likelihood that a person will start his own business. "^^ 

Managerial Experience 

Unlike Vivek and Humphrey, Barry had only a two-year degree. But he had more than a decade of 
experience as a manager of large and small teams and had had profit-and-loss responsibility for a 
large organization where he was a general manager who had to knit diverse functions into a coherent 
whole. As a merger-and-acquisition manager at GTE, Barry had regularly analyzed startup 
technologies and was practiced at identifying the strengths and weaknesses in new conpanies' 
business plans. Working through several boom-and-bust business cycles had given him the ability to 
maintain a strategic long-term view in the midst of uncertainty. 

Barry credits his prior managerial experience (and an MBA earned years later) with providing him 
with the tools to build and lead a team at his startup, Masergy. When his board differed with him 
about whom to hire for his executive team, Barry's experience gave him the confidence to stand his 
ground. "Less-experienced CEOs would have had a hard time pushing back against the board and the 
board's industrial psychologists, but I felt comfortable doing so," he explained. However, such pre- 
founding experience is relatively rare. Among the founders in my dataset, only 18% had management 
experience before founding their startups, including 19% of technology founders and 15%) of life 

sciences founders. (This is consistent with another study that found that technical founders tend to 
lack prior managerial experience and may even lack interest in developing managerial skills. 
Among the Inc. 500 list of the fastest-growing private companies, "the founders also often lack deep 
management or business experience. "^^) 

Functional Backgrounds 

Founding a startup requires the knitting together of all of the functions required to make an 
organization run effectively, fi-om product development to marketing to sales to finance to human 
resources. Having prior experience in those functions arms the founder with the ability to understand 
how each one operates on its own and as part of the larger whole. Founders who lack experience in a 
function either will be blind to key aspects of how it contributes to the whole or will have to spend 
valuable time learning about it rather than building the startup. One challenge for young founders is 
that it takes many years to accumulate multifunction or general management experience. For Barry 
Nails, it took many years of working at GTE to accumulate that type of experience, even though he 
was a high-performing employee there. 



Founders who have gaps in their functional backgrounds can be blindsided by problems in those 
areas. When Jim Triandiflou and his cofounder, Mike Meisenheimer, founded Ockham Technologies 
to offer a sales-force-optimization software product, both had strong sales and consulting experience. 
They understood their customers and the value proposition of their product, but lacked any 
background in software design and development. Jim recalled, "We had no idea the level of detail 
that you needed to write software. No idea. ... We used to joke that we'd have to take classes at 
DeVry to learn how to code software, because we don't [know how]." This inexperience caused 
serious delays and missteps with product development and prototype demonstrations. "Although 
consulting was brainpower-intensive, software was operationally [i.e., execution] intensive. In 
consulting, you do not get into nearly the level of detail that a real operation does. ... I look back 
now at some of my ideas and say I didn't have a clue. I was so unrealistic about what really gets done 
and how hard it is to get these things done," explained Jim Barry, too, found that his sales and 
product management background put him on a different timetable than his operations staff, making him 
impatient for faster progress. 

An executive's ftmctional background can also have a powerftil effect on the company's strategy 
and focus. In particular, when company executives have backgrounds in "output ftmctions" such as 
marketing, sales, and R&D, the company tends to emphasize product innovation, related 
diversification, and advertising. Conversely, when company executives have backgrounds in 
"throughput ftmctions" such as production and process engineering, the company tends to enphasize 

automation, up-to-date plants and equipment, and backward integration.^^ 
Industry Knowledge 

How well the founder understands the startup's industry can make a big difference in the challenges 
he or she faces; specific knowledge of an industry can help a founder avoid potentially fatal 
problems. Even before Barry Nails came up with the idea on which Masergy was based, his deep 
industry experience helped him play to his strengths. With his 25 years of relevant experience, he had 
developed enough human capital in the industry he chose for his startup to be his own in-house expert 
on potential target customers, what those customers would find desirable, and what his value 
proposition could entail. "As I started to build a new business plan, my planning was very similar to 
building a business plan for a product rollout at GTE. I asked myself a series of basic questions: 
'What do I know?' I know telecom, so the business should be in that. 'Whom do I know?' Enterprise 
clients. 'What do these enterprise clients care about when it comes to telecom? What are they willing 
to pay for?' " Barry knew how to answer these questions and he built a solid business plan around 
those answers. 

Vivek Khuller, on the other hand, had an industry- changing idea for online ticketing, but knew very 
little about the venues and events that would use such technology. "We were engineers who had never 
worked in this area. I've been to the Fleet Center [a Boston arena] just once. I have never been to 
Foxboro Stadium [a New England football stadium] to see a game and I've never entered [New 
York's] Madison Square Garden in my life. . . . [We're] not sports fans or anything of that sort." 
When Vivek first met with the operations manager of a large stadium, he realized he was woeftilly 
unprepared. "I had no idea how these venues were organized. . . . [T]his guy was concerned about 
how much more operational risk would he have to undertake by adopting this new technology. ... We 
had done very little research on that. We had very few answers to his questions." 



According to Professor Scott Shane's survey of the research on small businesses, "More than half 
(55 percent) of entrepreneurs start businesses in industries other than those in which they had 
previously been working."^^ Many of those founders like to believe that ignorance of the industry in 
which they are founding is a benefit because it leaves an opening for fresh thinking.^^ This can be 
true, but the advantages of ignorance are often easily outweighed by the disadvantages of 
inexperience. Past research has suggested that founders who launch startups in industries in which 
they haven't worked raise less capital, have lower enployment growth, and have a higher rate of 

failure than founders with prior industry experienced^ This should be particularly true when the 
mental models developed in one industry do not transfer smoothly to the new industry. For potential 
founders who have worked for a long time in one industry, and may even have had a lot of success in 
that industry, switching industries often requires substantially adjusting their mental models. Even 
when they do realize that they will have to adjust their models, they can be blindsided by parts of 
their models that unexpectedly cause problems or by industry- specific assumptions they didn't even 
realize they were making. 

For instance. Curt Schilling, a prominent former professional baseball player, would seem to have 
had tremendous advantages as an entrepreneur. One would expect that his status as a baseball star 
would open doors closed to others; that people would line up to work for him, advise him, and fund 
his company; that the financial capital he had accumulated during his baseball career would help 
launch his company; and that his work ethic and leadership abilities would be of great value in the 
business world. Some of those advantages did benefit Curt when he launched his MMOG startup, 38 
Studios. In other areas, however, he was in for a rude awakening, despite having spent some time 
observing the employees at a large gaming company. "There were so many things I didn't 
understand," he admitted. 

Some of those disconnects were to be expected for any first-time founder, especially one who 
hadn't yet worked full-time in the industry, but many could also be traced to the "career imprinf 
Curt had taken from his career in baseball and the mental model he had developed of how 
organizations work and how people are motivated. For example, he was initially mystified as to why 
people couldn't work 14 days straight and did not understand why they would need weekends off, or 
even regularly scheduled time away from the business. He had to adjust to how people were 
motivated and how teams were constructed in business. Baseball players are paid salaries and do not 
receive equity stakes in their teams, but employees in the MMOG industry expect to receive both 
salary and equity. Even seemingly innocuous terms he had taken for granted took on opposite 
meanings. Finding himself confused during a meeting, he finally realized that when businesspeople 
use the term "burn rate," they are referring to the amount of cash being used up and they want to keep 
it to a minimum For a baseball pitcher, "burn rate" is the speed at which you can throw the ball — the 
higher the burn rate, the better. After grappling with each confusion. Curt struggled to figure out his 
disconnect and begin adjusting his mental model, only to find further disconnects between his old 
mental model and his new industry. 

Working in Small versus Big Companies 

Working for someone else's startup helps you develop tacit knowledge about entrepreneurship that 
increases the likelihood of becoming a founder yourself People with more entrepreneurial 
experience have more ways to recognize and evaluate potential opportunities — more cognitive rules 



of thumb for knowing what to try and how to try it. Because of the division of labor in society, every 
person has different prior knowledge. To discover entrepreneurial opportunities, entrepreneurs must 
combine their prior knowledge with the right new knowledge, then figure out the correct combinations 
of means and ends with which to make the most of the opportunity.^^* 

The origin of Smartix is an example of such a combination. Mvek had a degree in electrical 
engineering and went on to work in a technical job for five years before entering business school. 
During his business school internship, he came in contact with a company that had developed 
electronic access cards. The internship, in conjunction with his prior education and experience, 
sparked a new idea: Vivek realized immediately how keycard technology could transform ticketing, 
allowing venues to ticket their own events. He was able to tap multiple elements of his background in 
order to recognize the technology as a particular kind of business opportunity. 

Learning about relevant patterns and causal relationships may be more likely to occur while 
working in entrepreneurial organizations. Employees of small companies are indeed more likely than 
employees of large companies to leave and become founders of their own companies. Also, 
companies that were once VC-backed were 20% more likely to spawn at least one startup than were 
companies that had never received VC funding. Another study found that half of all founders came 
from companies with fewer than 25 employees and 64% came from companies with fewer than 100 
employees, more than three times as much as would be expected since those companies account for 
less than 20% of all employees. One possible explanation for these results is that larger 
organizations maybe better at incorporating their enployees' innovative ideas and thus keeping those 
employees within the organization.-^^ Another possible explanation is that younger organizations may 
attract more employees who are entrepreneurially inclined. However, career expert Dr. Tim Butler 
cautions against gaining work experience in too immature a startup: "A more developed startup is a 
better training ground than two guys in a garage. A startup that is more mature, and has brought in 
experienced executives who will be good mentors, will prepare you better. This moment in the 
development of a company is a 'sweet spot' for [someone] who wants both a startup environment and 
access to talented manager-mentors with proven track records . . . [along with opportunities for] 
working closely with a truly seasoned manager as tough day to day decisions are made." 

Still, some large companies can provide "entrepreneurial career imprints" that prepare future 
founders to make the leap. For instance, people who worked at Baxter, a global healthcare company, 
in the 1970s were well prepared to help found and build biotech startups because Baxter provided 
high-potential employees with "mini-CEO jobs" that helped them develop the capabilities, 
connections, confidence, and cognition required to found.^^ Indeed, almost one-quarter of the biotech 
startups that went public from 1979 to 1996 had a team member who had worked at Baxter. In 
contrast, people working at Baxter's competitor, Abbott Labs, worked in more specialized roles and 
developed a "functional career imprinf ' that was much less relevant to founding a biotech startup. 
Former Abbott employees were much less frequently found on the teams of biotech startups that went 
public. In other large companies, roles in new-product development or in launching offices in new 
countries can also provide solid preparation for future founders. 

Although GTE did afford Barry some Baxter-like entrepreneurial experiences, the fact that he had 
worked for such a large company for so long resulted in a few blind spots he had to overcome when 
founding his own business. His challenges stemmed fi-om the disconnect between his big-company 
experiences and the core demands of a startup: to attract outside funding, to manage a board of 
directors, and to manage the distinct challenges of hypergrowth. Sensing that there would be a gap 



between his GTE e?q)eriences and the cognitive capital he would need to become a successful 
founder, he first spent short stints in two startups, but those experiences were not enough to fill the 
gap. "With investors, I didn't understand the funding options," Barry explained. "I didn't know any 
way to go about fiinding other than VCs." When he presented to VCs, Barry found that his business 
plan, which mirrored the ones he had prepared at GTE, was too detailed, failed to highlight the most 
inportant handful of milestones, and lacked focus on the management team Whereas the GTE name 
and reputation had been the focus before, Barry eventually learned that, in startups, "[investors] really 
care about, number one, the team," requiring him to devote the biggest segment of his presentation to a 
discussion of the individuals involved. 

Once he obtained fiinding, Barry struggled with how to manage his newly formed VC-led board, an 
entity he had not encountered during his time at GTE. One surprise was the amount of time required to 
prepare for his initial monthly board meetings: a full 25% of each month. He said, "I knew that VCs 
had to see what was going on in the company and I figured they would need much more than what a 
governing body at GTE needed, but I didn't know what they would be looking for." 

At first, Barry thought of his board as a boss and spent hours following through on their directives. 
Later, he realized that he had much more autonon^ than a boss/enployee relationship would imply. 
"From a governance standpoint, board members give you suggestions and ideas, but don't have 
responsibility to make sure you do them," explained Barry. "Their suggestions don't have any weight 
if you don't agree." Barry also didn't immediately understand which things he should delegate and 
which he should not; like many other founders, he had a hard time delegating and giving up control of 
decisions. After presenting a real estate report at one board meeting, one of his board members said, 
"We need you to do other things. If you ever come back in here knowing so many details, we're going 
to fire you!" Barry recalled. "That was a key message about their expectation of the CEO and what I 
should be doing as CEO. I was used to working hard, but it made me stop and think about how much 
to delegate versus how much I should take on myself" 

Barry also came to understand that the pace of growth in a startup created different management 
issues than the ones he had faced at GTE. "Prioritization is even more important [in a startup] than in 
a stable business," said Barry. "I realized that I needed people to not only prioritize what to do, I 
needed them to create not-to-do lists. It was easy for people to find new products for us to bring to 
market, new pieces to add, new customer segments to go after. . . . When you're smaller, if you go 
after something, it takes precious resources. You're also moving a lot faster, so you can harm the 
organization. . . . There's much more at risk, much more damage you can do." 

Harnessing Nontraditional Experiences 

We typically focus on schooling and work e?q)eriences as sources of human capital. However, 
founders can also gain relevant human capital outside of school and work, often in ways they did not 
anticipate. For instance, Tim Westergren's experience leading a rock band taught him leadership and 
management skills that he put to use founding the pioneer online radio service Pandora Radio. Tim's 
efforts as a conposer had given him invaluable sales e?q)erience and had taught him to persist through 
the ups and downs of searching for success, which proved to be a valuable commodity after he 
became a founder. Dick Costolo, the founder-CEO of FeedBumer, was a standup "inprov" comedian 
before he started his first IT conpany. Such a career path wouldn't seem particularly helpful for a 
budding entrepreneur, but in fact Dick learned quite a bit about human nature and team construction 
from his years on the stage. Reflecting on his inprov comedy years, Dick said, "My biggest team 



lesson was that you can't build [an improv] team from parts that look like they'll make a good team. 
We had a show with three people, all of them all-stars. Individually, they were hilarious and some 
have gone on to greatness. . . . But everyone could see they weren't working well together. ... [I 
learned] the fit between personalities was so much more important," a lesson he then applied when 
building his startup teams. 

Building Social and Financial Capital 

While people are building relevant human capital, they are often also building the social capital they 
hope to leverage as founders. In this context, social capital is the durable network of social and 
professional relationships through which founders can identify and access resources. Here, as with 
human capital, youth is often at a disadvantage. Barry, for example, spent years building connections 
with potential enployees, customers, advisors, and investors before he started Masergy. That deep 
reservoir of resources enabled him to meet his ambitious deadline of six months to make Masergy 
viable. Conpared to Barry, Vivek and Humphrey had narrow social fields. Vivek, however, was able 
to tap his past work contacts and school connections to build Smartix; his software developer was a 
former coworker at Bell Atlantic, and he recruited cofounders and advisors from his classmates and 
professors at business school. His school connections led to potential customers (Fleet Center, 
Madison Square Garden), which in turn built cachet to attract venture ftinding. All the same, he lacked 
the industry- specific connections necessary to fill the biggest hole on his team — the lack of an 
industry veteran. Barry's years at GTE had provided him with a wealth of exactly those types of 
connections. 

Financial capital is another area in which Barry had an advantage over Humphrey and Vivek. 
Many ftiture founders, while still working as employees, try to accumulate a cash cushion that will 
keep them going after leaving to work on the startup. The size of this cushion can largely determine 
the amount of time the founder is able to give the startup, the amount of stress and urgency he or she 
feels to become cash-flow positive, and the decisions he or she makes to build the startup. One 
founder described how he first had to attain a "walk-away" level of savings — enough to feel secure 
walking away from his current job to found a startup. Barry left his last company not only with his 
own savings but also with a six-month severance package, a cash cushion that both Humphrey and 
Vivek lacked but that gave Barry time to get Masergy off" the ground. In contrast to Barry and others 
who have accumulated financial capital, Humphrey and Cecilia had to live with his parents after their 
wedding in order to keep their expenses low enough to pursue ConneXus, causing considerable stress 
in their relationship and putting pressure on the startup to succeed quickly. Even so, Humphrey was 
more fortunate than many potential founders, who aren't able to make the leap at all for lack of 

financial capital. -^-^ One study found that 51.3% of people who had seriously considered becoming 
entrepreneurs could not do so because they lacked the necessary financial capital. 

The accumulation of one type of capital can spark a virtuous cycle. Research has shown that people 
who accumulate more social capital before founding are able to attract more human capital (such as 
cofounders) and financial capital (such as seed capital) with which to launch the startup,-^^ and to do 
so more quickly. 



The Perils of Waiting 



Although it is tempting to tell potential founders that they should work a long time so they can build 
the capitals needed to increase the chances of becoming successful founders, there are dangers in 
taking this route. First, it may not be a good idea to continually accumulate more human and social 
capital. Not only are there diminishing returns, but it may also be counterproductive to building a 
successful startup. According to one study, startups survive the longest when founded by people with 
midrange prior work experience, estimated at 25 years. Founders with more than 25 years of work 
experience faced higher probabilities of startup failure than did founders with approximately 25 years 
of experienced^ Second, by working for a long time, founders may get handcuffed to their pre- 
founding positions and may also become less fit to become founders. 

Stronger Career Handcuffs 

If people accumulate relevant capitals as they gain more work experience, they should become more 
likely to found a startup the longer they work. But as we saw at the beginning of the chapter, this isn't 
so. The likelihood of founding a startup does not increase with age or years of work experience. 

A major reason is that waiting to found introduces its own challenges. Working for many years can 
strengthen the handcuffs tying the potential founder to his or her employer. These include 
psychological handcuffs, such as the social status of an impressive title or well-known employer, and 
"golden" handcuffs, such as a high salary or a vesting schedule that requires one to keep working in 
order to earn stock awards. One founder vividly admitted, "I had to get off the heroin drip of a 
salary." All of these handcuffs reduce the likelihood of becoming a founder by raising the opportunity 
costs of leaving the employer, lowering the relative attractiveness of launching a startup, and 
reinforcing the inertia of remaining an employee. 

There may also be legal handcuffs, such as a noncompete agreement preventing the employee from 
competing with the employer (these tend to be particularly relevant to key senior employees) or an 
employer's claims on intellectual property developed by the employee while working for the 
employer. Such agreements made Jim Triandiflou, founder-CEO of Ockham Technologies, think 
twice about leaving his consulting job with The Alexander Group (TAG). Having become part of the 
senior management team, Jim found the appeal of going off on his own was somewhat darkened by 
noncompete and nonhire agreements that would prevent him from using a strategy similar to TAG's 
and even from taking on independent consulting projects to generate cash flow for the startup. 

Lower "Fitness to Found" 

Working as an employee for a long time can decrease not only a person's willingness to become a 
founder but also his or her fitness to do so. As people rise in their organizations and in their careers, 
they tend to develop specialties in which they have deep knowledge and contacts. As described 
above, depth of knowledge is important, but people who have depth without breadth do not become 
the sort of jack-of- all-trades that a founder needs to be.-^^ At GTE, Barry Nails avoided becoming 
specialized by working in a variety of positions, spending an average of less than two years in each 
position. Also, given the division of labor within formalized organizations, employees get used to 
relying on organizational infrastructure, processes, and support fiinctions, making the most eflRcient 
use of their own time but diminishing their own self-reliance. On both of these dimensions, a founder 
is the antithesis of the experienced executive and the transition to founderhood is made harder by the 



habits and mindsets that the experienced executive has developed. 

The when-to-leap decision can certainly be affected by the would-be founder's perception of the 
risks involved and how that perception changes over time. For instance, \lvek Khuller reflected on 
his decision to found Smartix: "We were all young and if, for any reason, it didn't work out, we each 
had lots of other options. ... At the same time, as you get older, you start getting a lot more cautious, 
prudent, maybe gun-shy. With the amount of energy it takes, you cannot do more than a few startups 
well. It just tires you. The amount of energy it takes from you, once you have a family and everything 
else, you don't want to deal with it." 

Stronger Family Handcuffs 

Single, childless founders are much less constrained in their founding decisions; the "family 
handcuffs" often tighten as spouses or children enter the picture. At Ockham Technologies, although it 
was Ken Burows who had come up with the idea for the startup, he was hesitant to join full-time after 
the birth of his first child. "When push came to shove. Ken . . . decided he didn't want to quit his 
job," explained cofounder Jim Triandiflou. "He had his first child. He just said it wasn't the right 
time for him. Which was the irony of all ironies, because he was the entrepreneur of the group. "^"^ 

Even potential founders with older children may be constrained in their founding decisions. For 
Barry Nails, both his emotional attachment to Texas, his home state, and the practical considerations 
of having a special-needs child led him to focus on founding in North Texas instead of heading to a 
more promising area for an entrepreneur. He said, "I absolutely did not want to move out of Texas. 
That was my limiting factor. I love Texas, Texas is home. It's where all my family is. I would move if 
I had to, but I really didn't want to. . . . [Also,] I had significant pressures to stay in North Texas 
because of all of the resources there for my autistic son: schools, therapists, doctors. Instead of trying 
to change that ... I made the decision that I was going to stay put and try to make things work in North 
Texas. I wouldn't do anything that would detract from our ability to support his requirements." 

Worried about having support at home, Barry and Vivek negotiated time frames with their spouses 
before starting their businesses; Barry received a commitment of six months and Vivek and his wife 
agreed to a year. While these constraints were challenging, the blessings and commitments of the 
spouses were crucial. "What they don't teach you in business school is to make sure your life partner 
is in sync with what you're doing," Vivek explained. "You can divide your 24-hour day into three 
parts: eight hours work, eight hours personal, eight hours sleep. If your eight personal is not in sync 
with your eight work, then your eight sleep will suffer." 

In contrast, as Humphrey desperately tried to secure funding for ConneXus, his wife and parents 
hoped he would soon leave the startup and pursue a more secure job. Their lack of support made an 
already difficult situation even harder. He explained, "You have to make sure that your family and 
your spouse, all the people you love and care about, are supportive of what you are doing. There is 
always a temptation to abandon the idea, to jump ship. If you don't have the support around you, it's 
much harder to resist. If you don't have 100% support, wait until you do, and [start a company] then." 

Part of the problem that Cecilia, Humphrey's wife, had with ConneXus was Humphrey's 
unrealistic timeline, which she realized only in retrospect. Originally, he told her that he would 
secure funding by the time they were married in October. But by the end of the year, the couple was 
still living with his parents and the startup had not yet received the promised funding. If Humphrey 
had painted a more conservative — or at least realistic — scenario for Cecilia, rather than trying to 



"sell" her on letting him found ConneXus, she might have been prepared for the uncertainties of the 
startup process and the down parts of the entrepreneurial roller coaster. But as another founder said, 
"My wife doesn't know that it will probably take twice the time I expect and cost twice as much as I 
expect. If I tell her that, will she even let me try?" 

An entrepreneur may also be constrained by his or her personal financial situation — the mortgage, 
college loans, credit card debt, and so on. A working spouse can be crucial. When Tim Westergren 
got the idea to start Pandora, he was entering his 30s and knew he wanted more financial stability. 
The fact that his wife's salary provided them with a financial buffer enabled him to go ahead with his 
startup. A survey of founders of high-growth businesses found that 70% were married when they 
founded their startups and that 60% had at least one child.^^ For these founders, a nonworking spouse 
makes the challenges particularly acute; research has shown that people whose spouses were not 
employed were less likely to become entrepreneurs.^^ 

Developments That Release the Handcuffs 

Although career handcuffs, golden handcuffs, and family handcuffs tend to get stronger as time passes, 
other developments can release these handcuffs and make it more possible for the potential founder to 
launch a startup. These developments, which are summarized in Figure 2.2, include the following: 

An Employer's Slowing Growth 

Employees tend to leave companies to become founders when their employers' growth rates begin to 
fall, reducing the attractiveness of staying there and the opportunity costs of becoming founders."^^ The 
likelihood of an employee leaving to become a founder has also been linked to the enployee's own 
performance. Employees whose performance was in the middle of the distribution were the least 
likely to leave; the most likely were the poor-performing employees (the "slugs" whose pay wasn't 
high and who had the least to lose by leaving to found a startup) and the high-performing employees 
(the "stars" who had the potential to earn high wages by becoming self-employed).^ Although high- 
paid employees may be less likely than lower-paid employees to walk away from their salaries, if 
they have saved a high percentage of their earnings over the years, that nest egg may make them more 
likely to make the leap than if they hadn't saved. 

An Employer 's Change in Strategy 

The employer's acquisition or change of strategy may spark the potential founder's leap. Barry 
Nails 's work at GTE was consistently interesting enough to keep him there, his startup dream 
deferred, until the firm announced its impending merger with Bell Atlantic, another large telecom 
conpany. Barry explained why the merger pushed him to leave: "GTE had a lot of freedom; we could 
do a lot of innovative things. But when we started the merger discussions with Bell Atiantic, it was 
apparent that we were so different in culture, in the degree of innovation, and in the ways we worked. 
And there was the probability that my job would be moving to the Northeast, which I didn't want to 
happen." 



Employment Shocks 



For founders, probably the most drastic career trigger is the loss of a job. Genevieve Thiers was 
working for the Lotus division of IBM while developing her babysitting website, Sittercity, on the 
side. She had been planning to leave IBM eventually to work full-time on Sittercity, but in 2002 the 
economic downturn caused IBM to lay off the entire Lotus division with a six-month severance 
payment. As Dan Ratner, Thiers 's husband and early employee, explained, "The biggest inflection 
point for an entrepreneur is when to quit the day job. Genevieve was fortunate enough to have the 
decision made for her." 



Handcuffs 


Developments That Can Release the Handcuffs 


Career handcuffs 


Employer's growth slows 




Employer changes strategy or is acquired 




Being laid off or fired 


Golden handcuffs 


Severance payment from layoff 




Inheritance or large gift 


Family handcuffs 


Spouse starts working 




Children grow up 




Emigrate to country (or region) that fosters 




entrepreneurship 



Figure 2.2. Developments That Release Each Type of Handcuff 

Financial Shocks 

Triggers can also be positive. Unexpected personal windfalls, for example, can spark the decision to 
leap; people who receive an inheritance or a large gift are more likely to found businesses. 
(Negative financial shocks can also alter the founding equation, but usually by precluding or delaying 
the decision to found.) 

Family Developments 

Although Humphrey Chen's family developments hindered his leap, family developments can also 
trigger a leap. These triggers can range fi"om marrying someone whose paycheck can support a leap 
into founderhood to having one's children become more independent. For Robin Chase, founder of 
car- sharing company Zipcar, having her young children start school offered the fi"eedom to launch her 

startup. Moving to a country or region that is more supportive of entrepreneurship may also release 
the potential founder fi*om familial or cultural handcuffs. 

HOW CAN I DISPASSIONATELY EVALUATE MY IDEA? 

To ensure that the best decisions are being made about whether and when to launch their startups, 
potential founders should also be evaluating as objectively as possible the potential ideas on which 
they might base them. However, a founder's natural passion and confidence can prevent an objective 



evaluation and set up the founder to fail. 
Evaluating Ideas 

A study of fast-growing startups found that, for 71% of them, the founder got the idea while working 
at a regular job. Tim Westergren's idea for Pandora Radio was sparked by his work as a musician 
and composer. For other founders, the idea for a startup may feel like it came out of the blue, while 
actually being grounded in some aspect of his or her past. For instance, when Genevieve Thiers 
happened to encounter a pregnant woman looking for a reputable babysitter, an inspiration hit her: 
"Wouldn't it be interesting to put a listing of all of the babysitters in the country in one place?" In fact, 
her own experience as a teenage babysitter helped spark that business idea and gave her an instinctive 
sense that it would fill a broad and important need. (Similarly, Vivek Khuller's encounter with a 
conpany that had developed electronic access cards inspired his business idea, but this idea was also 
sparked by his experience as an electrical engineer.) 

Other founders are more like Barry Nails, who had to work to discover an opportunity. As 
described above, Barry had to take a top-down approach to finding an appropriate industry segment, 
specific target customers within that segment, and then the services for which they would be willing 
to pay. In going through that process, Barry's years of experience both selling to customers and 
learning the telecom industry were the foundations on which he could build the idea for Masergy. 

Regardless of how the idea was generated, a key step in the decision to leap into founderhood is 
the evaluation of the idea's potential — whether the "market window of opportunity" is favorable. 
Academic research has emphasized the value of market analysis by establishing connections between 
industry characteristics and startup success, \blumes have been written about how to analyze the 
attractiveness of industry and market opportunities and readers who plan to do such an analysis are 
referred to those volumes. The following are some of the questions that potential founders ask about 
the market window of opportunity and how some of the academic research provides guidance: 

Market potential: Are customers willing to pay for such a product or service? How big is the 
market? Is it growing? Startups with a broader potential reach (e.g., they can target a national 

market instead of a local or regional one) have higher survival chances. Startups with products that 
"disrupf ' the market by introducing a new architecture or component are more likely to succeed than 
those whose products are more "sustaining" and thus compete head-to-head with those of established 

competitors.^^ But startups with disruptive products may also have to plan for more development time 
than would be needed for sustaining products, for which a market already exists. When Vivek Khuller 
gave himself one year to establish a viable company, he did not realize that, for an innovative, 
industry- changing product like Smartix's, one year was probably not enough. 

Competitive landscape: Is it favorable? Are many companies competing for scarce 
resources? When a groundbreaking company is founded and creates a new niche in the market, it 
does not have to compete with other players for resources and customers. But as more conpanies 
enter that niche, the competition increases rapidly until the population of companies outgrows its 
market niche and disbanding rates become greater than founding rates. The result is an S-shaped 
pattern of growth within the industry niche, with the early stages of growth being much more attractive 

competitively than the later stages. At the points on the curve where the niche is more "munificenf 
with plentiful resources available to companies, organizational birth rates are highest.-^^ Startups in 



highly competitive environments have lower survival chances than in lower-conpetition 
environments.^^ 

"Ticking clock": Is there a ticking clock that requires me to move quickly to pursue my idea? 
Is the window of opportunity about to close? Ticking-clock industries include (a) those in which 
products and services are quickly derived from and just as quickly outdated by technological or 
scientific advances, (b) those with strong "network effects" — that is, the value of the product 
increases as more people use it,^^ and (c) those with significant economies of scale. Such industries 
tend to have stronger first-mover advantages, while in other industries, late entrants have a decent 
chance or may even have an advantage — a chance to outdo the first movers by learning from their 
mistakes and allowing them to develop customer awareness and a supply chain. 

For the eventual founders of Proteus Biomedical, a developer of miniature computers and sensors 
for life science applications, understanding whether or not they were facing a ticking clock was 
critical. In 1989, Andrew Thompson and George Savage, who had degrees in engineering and 
medicine, respectively, and had met at Stanford Business School, saw an article in Time magazine 
that described micro-electrical mechanical systems (MEMS). Andrew envisioned these as "little 
scissors that one day would fioat around in your bloodstream, gobbling up cholesterol" and the duo 
was intrigued by the possibility of a startup. They concluded, however, that "while applications like 
these were potentially interesting, they couldn't be built yet. We shelved the idea, but periodically 
asked ourselves, is it time?" George said, "Healthcare generally isn't a good place to try out 
revolutionary technology that hasn't been proven elsewhere. "^^ A decade later, however, Andrew 
and George decided that the technology to develop MEMS had matured to the point where the market 
for it might emerge quickly, sparking their founding of Proteus Biomedical. 

Startup ideas must also be evaluated in light of the specific person's capabilities and skills. The 
same telecom opportunity would be seen in a very different light if it were being pursued by Tim 
Westergren instead of Barry Nails, or even by an early-career Barry Nails instead of a late-career 
one. A potential founder who has accumulated the right capitals should have very different answers 
about whether and when to leap than one who has not done such a good job preparing for the leap. 
For instance, a well-prepared founder should have more confidence in his or her ability to quickly 
create and deliver a product worth paying for and to identify and convince potential customers. 

Clouding Judgment: Passion and Optimism 

Passionate founders who have evaluated all of these dimensions of the opportunity must still guard 
against their natural overconfidence and optimism, which may have skewed their evaluation. 
Founders tend to be highly confident in their own abilities, in the potential of their ideas, and in the 
prospects of their startups. In a survey asking entrepreneurs to compare the prospects of their startups 
to those of similar startups, 95% of the respondents believed that their own startups had a better than 
50% chance of succeeding, but only 78%) believed that a similar startup had the same chance of 
succeeding. A fiill one-third of the entrepreneurs believed that their startups had a 100%o chance of 

success. The authors of that study concluded that entrepreneurs tend to be overconfident in their 
prospects. In another study, entrepreneurs were found to be 20%o more overconfident about their 
ability to answer factual questions about medical issues than were non- entrepreneur managers. Such 
overconfidence may encourage entrepreneurs to take action sooner than they should have or to have 
overly rosy views of their prospects. 



Is optimism an advantage or a disadvantage for a potential entrepreneur? On the one hand, 
optimism may spark more innovation and less of a tendency to "follow the herd."^^ Founders who are 
optimistic tend to act more quickly and to launch faster- growing businesses. However, optimistic 
founders also tend to create unrealistic business plans based on rosy projections and to underestimate 
their competition.^'^ They are more likely to overcommit to the initial idea instead of flexibly 
adjusting it and to underestimate the resources they will need to build the startup, thus increasing their 
chances of failure. 

Multiple studies have found entrepreneurs to be more optimistic about the future than non- 
entrepreneurs.^^ For serial entrepreneurs, that can actually be a disadvantage, leading them to 
discount negative information, to be more susceptible to confirmation biases,* and to repeat their past 
actions rather than question whether those actions apply to the current opportunity. As a result, higher- 
optimism entrepreneurs have 20% lower revenue growth and 25% lower employment growth than 
lower-optimism entrepreneurs, who would be less susceptible to the perils of optimism. 

In short, optimism is often a double-edged sword. Venture capitalist Guy Kawasaki, in his blog 
"How to Change the World," lists the "top ten lies of entrepreneurs." Most of these lies have to do 
with founders' oversized confidence and naive expectations, fi^om their projections of revenue and 
market size to their "proven" management team to their lack of competition. When an entrepreneur 
says he or she will make $50 million in four years, explained Guy, "I add one year to delivery time 
and multiply [the revenue] by .1."^"^ 

Founder optimism can also affect the decision about when to leap, leading some to leap sooner 
than they should and others to leap when they shouldn't at all. As we will see throughout this book, 
overconfidence and optimism also play a role in many subsequent founding dilemmas, from the 
decision about whether to found alone or seek the help of cofounders, to the financial terms a founder 
is willing to accept from investors, to a founder's expectations about how long to remain CEO of his 
or her own startup. 

CLOSING REMARKS 

One hears a lot about "following your passion." Potential founders should avoid the mistake of 
thinking that their passion excuses them from a rational assessment of their circumstances. The French 
author Fran9ois de La Rochefoucauld observed that "[t]he heart is forever making the head its fool," 
and Steve Jobs, cofounder of Apple Computer, was fond of saying, "Follow your heart, but listen to 
your head." Sometimes the head is right and the heart is wrong, at least for the time being, and 
especially for passionate overconfident founders. To make sure that "the head" is part of the decision, 
potential founders need to first ask the three big career questions examined in this chapter and 
summarized in Figure 2.3. 



Should I become a 
founder? 

I Yes 



No 



Remain nonfounder 
(or do part-time); 
re-evaluate later? 



When should I 
found? 




Dispassionate 
evaluation of idea? 



Act to improve 
circumstances, 
then re-evaluate 



\ 



Favorable 
personal, career, and marlcet 
circumstances? 




Evaluate other 
Ideas 



No 



No 



Yes 



Leap into founderhood! 



Figure 2.3. Three Big Career Questions 

Humphrey's and Barry's decisions about whether, when, and what to found were powerfully 
affected, in very different ways, by where they were in their careers, by the characteristics of their 
startups, and by their personal situations. Figure 2.4 shows the three recurring factors we have 
examined — career factors, personal factors, and market factors — that affect the three questions 
outlined in Figure 2.3. 

When founders see all three factors as being favorable — that is, if the situation falls into the small 
black "bull's-eye" region at the center of Figure 2.4, where the founder's career experiences provide 
a solid foundation on which to build the startup, his or her personal situation is conducive to 
becoming a founder, and the market is favorable — then it's time to follow the heart and make the leap! 




•Relevant work experience, mental model 
•Broad/deep wori< experience 
♦Low opportunity costs 
•Change in golden handcuffs 
•Fitness to found 



•Entrepreneurial motivations 
•Supportive family situation 
•Positive role models 
•Cash cushion 



•Big opportunity 
•Favorable context 
•Customer willingness to pay 
•Ticking clock 



Figure 2.4. Career, Personal, and Market Circumstances 



However, many founders suggest that such a perfect alignment of all three factors is rarely, if ever, 
encountered; there is no "perfect time" to become a founder. Potential founders who keep waiting for 
perfect alignment may come to the end of a long career regretting that they never made the leap. 
Nevertheless, founders who are overly optimistic or who misread their circumstances as being more 
favorable than they truly are will be taking unnecessary risks by leaping into founderhood when the 
three factors are really not favorable. Like those who waited for the perfect moment that never came, 
these overconfident potential founders may also end their careers with remorse, feeling that they 
squandered too soon the resources that might have made for a more successful leap later on. A better 
way to proceed when the factors are clearly not favorable is to dispassionately evaluate which 
circumstances need improvement and then to act to improve them prior to launching the startup. 

Thus, the toughest dilemmas are the much more common situations in which the potential founder is 
passionate about becoming an entrepreneur but only two of the three factors are favorable. If 
perfection is never to be expected, how close is close enough? These situations are represented by 
the three gray regions in Figure 2.4. Each region poses its own challenges for the potential founder, 
but each challenge comes with its own set of possible solutions, summarized in Figure 2.5 below. 
However, as potential "gray-region" founders tackle the remaining area of weakness in an effort to 
move into the bull's-eye, they need to remember that they are aiming at a moving target and to make 
sure that other factors have not become unfavorable. Let's examine each gray region: 



"Gray-Area" Obstacles Potential Solutions for Getting to the "BtiU's-Eye" 

"I have everything but Top-down evaluation of potential markets, 

an idea." customers, and business models 

Find "idea founder" whose own holes match your 
strengths 

Participate in other startups as an advisor or seed 
investor 

Attend conferences and read journals and blogs in 
your area of interest/expertise 

Find job to fill functional or industry hole 
Attend entrepreneurial-networking events 
Find complementary cofoundcrs, advisors, or 
mentors 

Fill hole by attending targeted class or program 

Probe spouse's fears and openly explore ways to 

address them 
Agree on realistic timelines and resources that will 

be committed to the startup 
Explore whether part-time founding is feasible 

Save money to supply cash cushion and seed 
capital 

Maintain low "personal burn rate" 
Resist accumulating redundant or irrelevant 

human capital and social capital 
Seek broad range of assignments in various 
finictions 

Scnitinize all noncompetes and other agreements 
before signing 



Figure 2.5. Obstacles and Potential Solutions for Gray-Area Potential Founders 

"I Have Everything But an Idea!" 

When Barry Nails decided to quit his job and start a new company, he was comfortable with his 
personal and career positions. Professionally, he felt he had gained the skills, contacts, and 
experience he needed — both from GTE and from his time at various early-stage companies — to create 
his own startup. On the personal front, his wife was frilly supportive. But Barry, unlike many of our 
other enfrepreneurs, had never had an idea hit him, even though he had a broad idea of the business 
segment in which he might find an idea. Without an idea, he had no "market circumstance" to even 
evaluate, let alone pursue as a startup. To solve this problem — two favorable factors but no idea — 
Barry sat down and very systematically developed his business concept using the process described 
earlier in this chapter, then built a business plan to take advantage of his abundant human and social 
capital. 

Another option for potential founders who lack an idea is to seek a cofounder who does have an 
idea but lacks something else that the idea-less founder can provide. For instance, James Malmo had 
business savvy and multiple ideas for high-tech products but lacked the engineering skills to 



"But I'm not armed for 
battle yet" (especially early- 
career founders). 



"But I'm going to hurt 
my family." 



"My handcuffs have gotten 
loo strong" (especially late- 
career founders). 



implement them. He began attending meetings of the MIT Entrepreneurship Club, where he met Javier 
Pascal, an MIT engineer in search of an idea, and the two eventually became cofounders of Lynx 
Solutions. Other sources of complementary "idea founders" include coworkers — as was the case for 
the founders of Ockham Technologies, who were coworkers at The Alexander Group — and school, 
where the Smartix cofounders met as classmates. 

"But I'm Not Armed for Battle Yet!" 

Tim Westergren had a supportive spouse and had already had an epiphany that led to his idea for 
Pandora, an online radio station; his personal and market circumstances were favorable and 
beckoning. However, Tim was reluctant to leap because he felt he lacked the professional skills and 
experience to tackle a startup. As a former musician and composer, Tim had no idea where to begin. 
"It was a complete mystery to me how you start a company. I still had not had a real job. I didn't even 
know the basics." 

This is often the situation for early-career founders. When planning their startups, such founders 
should assume that they are missing important skills and contacts and then carefully diagnose the gaps 
in their human capital, social capital, and financial capital. By attending a class or by finding a job 
that provides the professional skills that are lacking, inexperienced potential founders can try to 
develop the specific kinds of expertise they need. To fill social capital gaps, they can search out — 
and force themselves to regularly attend — local networking events, making sure to follow up with 
new contacts. Over time, regular networking can yield a much stronger set of contacts. 

Along the way, potential early-career founders should also be searching for potential cofounders, 
hires, and investors/advisors. The involvement of one or more of those players can help fill in the 
gaps in the various categories of capital, can provide the objective perspectives necessary to counter 
the founder's natural passion and optimism, and can provide guidance about decisions and patterns 
that aren't yet a part of the potential founder's mental model. For example, it wasn't until Tim 
Westergren's wife introduced him to Jon Kraft, who had founded a high-tech conpany in Silicon 
Valley and had the business skills Tim lacked, that Tim was able to fill his career holes and found 
Pandora. (However, as described in Parts II and III, founders should carefiilly weigh these benefits 
against the challenges introduced by involving other players in the startup.) 

"But I'm Going to Hurt My Family!" 

Although Tim and Barry eventually entered the bull's-eye region of Figure 2.4 — Barry by using his 
own reservoir of knowledge to come up with an idea and Tim by partnering with a key cofounder to 
gain the professional capabilities to start a company — Humphrey Chen's experience suggests that the 
most insurmountable obstacle to founding may be unfavorable personal circumstances. One cannot 
alter them so easily by taking purposefiil action or by seeking the help of others. Humphrey's market 
and career circumstances were favorable and he was passionate about building ConneXus. But his 
personal circumstance was quite unfavorable — although he didn't fiilly realize it — and about to get 
worse. His new wife craved stability and was risk-averse; both she and his parents favored his fiill- 
time consulting offer. When Cecilia became pregnant with their first child, Humphrey decided to take 
a secure job with Microsoft. 

The challenges of balancing startup demands and family demands can be particularly acute. While 



reflecting on his own experiences trying to achieve that balance, serial entrepreneur Steve Blank 
came up with "four big lies entrepreneurs tell themselves about work and family."^^ Those lies were 
"I'm only doing it for my family," "My spouse 'understands,' " "All I need is one startup to 'hit' and 
then I can slow down or retire," and "I'll make it up by spending 'quality time' with my wife/kids." 
He concluded that, for him, "None of these were true." Because of that, "Over time I began to 
recognize and regret the trade-offs I had made between work and relationships." 

A possible solution to an unfavorable family situation is to work on the startup part-time — after 
hours or on weekends — either as a long-term approach or as a transitional phase with the option of 
leaping full-time later. Proposing or agreeing to such an option may help persuade a spouse to support 
the entrepreneurial endeavor. However, in some situations, agreeing to go the "safe" part-time route 
can reintroduce the career handcuffs and can actually increase risks, slowing down progress and 
letting the market window slam shut. Mvek Khuller observed, "My wife could have been strict. She 
could have said, 'Go ahead with Smartix, but also have a part-time job so you can pay your share.' If 
she'd done that ... I would have gotten another job and would not have had the time or energy to do 
Smartix." 

In industries where competitive pressures force founders to move quickly, it is indeed advisable to 
try to dive into founderhood full-time. Where there is less rush to build the startup, founders can often 
manage their personal risks and gain buy-in from their families if they take a more gradual approach 
to founding. However, cofounders or potential funders may force the founder to make a full-time leap 
sooner than planned. Jim Triandiflou was still working at The Alexander Group and working on 
Ockham Technologies part-time when a potential angel investor, Bobby Crews, pushed him and his 
cofounders to quit their day jobs. As Jim recalled, "Crews said, 'We are interested in doing it, but 
you've got no customers, you haven't even quit your dang jobs yet, and you have nothing for a 
product. . . . Right now, you're playing both sides where you got a job and you're making money. 
We've got to see substance and we've got to see commitment.' " As a result, a month later, both Jim 
and his cofounder, Mike Meisenheimer, quit their jobs. 

"My Handcuffs Have Gotten Too Strong!" 

Potential late-career founders have to guard against the perils of waiting. They should avoid the trap 
of continually accumulating more human and social capital. They should also maintain a low 
"personal burn rate" so they don't get used to the handcuffs of a high-priced lifestyle and so they can 
accumulate seed capital for their eventual startup. 

In retrospect, many late-career founders regret not having made the leap sooner. Four years after 
founding Masergy, Barry Nails lamented, "Boy, I wish I had become an entrepreneur long ago! Why 
did I wait so long?" A year after founding a startup that provides customer-behavior analytics 
services, founder-CEO David Wellman reflected on his decision to leap:^^ 

Well today is my birthday, 40. And I am just finishing year 1 of my own venture. I have professional investment backing. A 
product in its infancy. Two large clients and a dozen smaller ones and a lot in the pipe. 

So that means I have that CEO title, and I was sweeping floors this morning. It's 3:30AM and I am still at work, for the second 
aU-nighter this week. I don't have enough people, time, or energy to deal with all the daily problems — like making a Costco run to 
get more printer paper. 

Yes, the hours are arduous, [and] my customers [are] more demanding than the most incompetent boss I have ever had. My 
pay is paltry and I sometime need to give it back. The work environment is less desirable than my college dorm room. 
But the second half of my career looks so good from here ! 



Resolving the career dilemma in favor of making the leap is a big and exciting step, but it leads 
directly to many more dilemmas, each with the potential to smother a young startup in its cradle. In the 
coming chapters, we will examine some of the most inportant — and most interesting — of those 
founding dilemmas, many of which are directly tied to the pre-founding career decisions discussed in 
this chapter. Founders who make the leap early in their careers should be more inclined to look for 
help from complementary cofounders, hires, or investors and should be more willing to give up the 
equity, decision-making control, and compensation required to attract them In contrast, late-career 
founders who already possess much of the capitals needed to launch the startup should be able to 
consider the option of being a solo founder and should also be able to attract complementary players 
by offering less attractive packages. Each of these career paths thus leads to very different potential 
outcomes. 



* As described in Chapter 1, the centrality of wealth and control choices is echoed by the (very different) Kauffman and Panel Study 
of Entrepreneurial Dynamics datasets. I emphasize the CareerLeader dataset because (a) it is much larger (its 27,000 executives and 
entrepreneurs dwarf past datasets that were used to examine entrepreneurial motivations), (b) it enables separate analyses for each 
gender, stage of life, and entrepreneurial status, and (c) it is based on rich underlying psychological research that uses multiple 
approaches to evaluate motivations. 

* Experienced entrepreneurs seem to be more likely to use "effectual reasoning," wherein they start with a given set of means — their 
personal strengths and the resources they already have at hand — rather than a predetermined goal, and then allow opportunities to 
emerge to which they can react. In contrast, non-entrepreneurial executives tend to use causal reasoning, in which they set a goal and 
then seek the best ways to achieve it. For more details, see Sarasvathy (2008). 

* A confirmation bias is a tendency to favor information that confirms preconceptions, regardless of whether the information is true. 
It may be caused by wishful thinking or by bounded rationality (ie., limitations on people's abilities to process information) and often 
contributes to overconfidence. 



PART II 

FOUNDING TEAM DILEMMAS 



INTRODUCTION 

The light bulb has gone off; a new business idea has sparked the decision to launch a startup, and the 
founder now faces a myriad of decisions. In this part of the book, I first focus on the dilemmas faced 
by the "core founder," the person who had the idea to launch the startup and is driving the initial 
stages of building it* The first of these dilemmas is the solo-versus-team dilemma: Do I begin the 
startup by myself or do I attract cofounders to help me start it? 

The decision to attract cofounders introduces further critical decisions about whom to attract, what 
role each cofounder should play, and how to split the equity among the cofounders. I call these three 
core dilemmas "the Three Rs" — relationships, roles, and rewards. In Chapters 4 through?, I use the 
Three Rs framework to examine how founding teams can, do, and should make these decisions; how 
these decisions are interconnected; and how they will affect the outcomes of team stability, company 
valuation, and control. 

OVERVIEW OF CHAPTERS 

The first chapter in this part will focus on the initial solo-versus-team dilemma. The subsequent 
chapters focus on team issues, with a chapter devoted to each of the Three Rs, followed by a chapter 
examining the linkages among them A brief overview of the chapters in this part follows: 

Chapter 3: Solo-versus-Te am Dilemma — Founding solo can help avoid the relationship, role, 
and reward dilemmas described below but introduces its own challenges and risks, such as a 
lack of human, social, or financial capital and a lack of emotional and psychological support 
during the toughest parts of the entrepreneurial roller-coaster ride. 

Chapter 4: Relationship Dilemmas — The cofounders' previous shared experience — or lack of 
it — affects the team's ability to tackle tough issues together, its diversity of skills and 
perspectives, and its stability, sometimes in surprising ways. 

Chapter 5: Role Dilemmas — The division of labor within a founding team can be a major 
source of tension if cofounders are fighting over titles, if they have overlapping roles, if the 
collective decision making fosters gridlock or interpersonal conflict, or if the division of 
labor cannot be adjusted as the startup evolves. 



Chapter 6: Reward Dilemmas — The division of equity and the allocation of financial benefits 
can cause problems within teams if members feel they are not being compensated fairly for 
their contributions to the startup. Compensation arrangements that align interests and protect 
against negative surprises are hard, yet critical, to accomplish. 

Chapter 7: The Three Rs System^ — When founding-team decisions are aligned across all three 
dimensions, the team is more likely to be stable and experience lower turnover. Conversely, 
misalignment is likely to lead to greater intrateam tension and higher turnover. Thus, to fully 
understand the factors that lead teams to splinter, it is critical to examine not only each R on 
its own but also their mutual alignment or misalignment. 

In these chapters, I describe each dilemma in detail, survey the relevant research, and discuss case 
studies of founders and founding teams to illustrate the interpersonal tensions and dynamics and to 
provide examples of best practices. Where possible, we will examine how two main outcomes, the 
founding team's stability and the growth in startup valuation, are affected by early decisions about 
relationships, roles, and rewards. Although we will examine each set of dilemmas separately, the 
dilemmas are interdependent and may be encountered in different sequences, a theme to which we 
return in Chapter 7. 

In each chapter, I will present data that show the pervasiveness of the patterns described in the 
field cases. In 2006, I enriched the founding- team section of my annual survey; thus, for this part of 
the book, I make extensive use of the data collected since then (i.e., a dataset that combines the 2006, 
2007, 2008, and 2009 surveys). This dataset includes 4,232 founders from 1,542 private startups in 
technology and the life sciences. Of the startups in this combined dataset, 88% were founded between 
1998 and 2008. Appendix A provides further details on the survey and the dataset. 

The data are complemented by case studies of founders who faced these dilemmas, sometimes 
successfully but often not. For instance, we will examine the decisions of Evan Williams, a serial 
entrepreneur who took a variety of approaches across his startups, often learning from his prior 
outcomes. In each chapter, we will meet Evan and several other founders from whom we will have 
much to learn. 



* As described in Chapter 1 regarding the Circles team, it is possible for a founding team to have more than one core founder. The 
dilemmas described in this part also apply — and sometimes apply even more strongly — to such teams, for these core founders must also 
decide how to manage the challenges introduced by their prior relationships, how to split the roles and decision making, and how to split 
the financial rewards among themselves. 



CHAPTER THREE 



THE SOLO-VERSUS-TEAM DILEMMA 



As WE SAW IN Chapter 2, career, market, and personal factors play a powerful role in the 
decision to become a founder. They also play a central role in one of the first and most important 
decisions faced by a founder: whether to "go solo" or to form a founding team. People who have 
thought carefully about whether to become founders and have considered the issues raised in Chapter 
2 gain insights that also prove useful for the solo-versus-team decision. In contrast, people who go 
solo when they shouldn't will increase their risk of failure, while people who bring on cofounders 
when they should have gone solo will usually face team tensions they could have avoided. 

As shown in Figure 3.1, in my dataset of high-potential startups, only 16.1% were solo-founded, 
with a high of 17.5% in the technology industry and a low of 11.7% in life sciences. More than one- 
third of the startups had two founders and one-quarter had three. What separates the minority of solo 
founders in these startups fi^om those who build founding teams? 

REASONS TO GO SOLO 

There are a number of reasons why a founder might choose to go solo. For instance, the founder may 
already have many of the resources needed to found the startup, or there may be little need for the 
startup to grow quickly. The startup may not be expected to grow big enough to support multiple 
founders. The founder may want to keep all the equity and control of all decision making, which 
would be particularly true for control-motivated founders. The founder may also be tempted to keep 
things simple in the early stages and thus dodge the coordination problems that come with bringing 
cofounders on board. Here, the founder's instinct (or prior experience) is backed up by Arthur 
Stinchcombe's groundbreaking article on what makes new organizations especially prone to failure.^ 
Three of the four factors he identified were repercussions of forming a founding team: cofounders 
were often challenged — and often undone — ^by the need to (a) learn new roles within the founding 
team, (b) negotiate the distribution of economic rewards among themselves, and (c) form trusting 
relationships with strangers (i.e., each other). 



45.0% 

40.0% 

35.0% 

^ 30.0% 
E 

S 25.0% 
^ 20.0% 
^ 15.0% 
10.0% 
5.0% 





1 


2 


3 


4 


5 


6 


7 


OTechnology 


17.5% 


39.0% 


21.9% 


12.2% 


5.5% 


2.1% 


0.6% 


■Life Sciences 11.7% 


34.8% 


29.1% 


11.7% 


8.4% 


1.5% 


0.6% 



Figure 3. 1. Sizes of Founding Teams for Technology and Life Sciences Startups 




With one or more of these reasons in mind, a founder might conclude that this is something he or 
she can do without cofounders, so why ask for trouble? Barry Nails, for example, founded Masergy, 
an enterprise telecom company, on his own, convinced that he had the resources he needed and that he 
was better off making all his own decisions. He had worked at GTE for 25 years, starting off with 
only a two-year technical degree and working his way up the ranks (earning an MBA along the way), 
gaining valuable experience managing teams, spearheading new programs, and running sales and 
marketing groups. In the mid-1980s, he had taken a break from GTE to start a small business, but it 
stagnated without adequate fiinding. After returning to GTE as a product manager for new 
technologies for several more years. Nails had again tested the entrepreneurial waters by joining a 
small startup in Texas, but soon left that startup after a disagreement with the CEO. By the time Nails 
decided to found Masergy, he felt he had the broad range of skills he would need. He had sales and 
operating skills, deep knowledge of the industry, contacts with potential customers, and a cash 
cushion to provide living expenses while he developed the business plan. He could wait to hire and 
involve other people, and in the meantime saw little to gain by adding a cofounder. Going solo made 
sense. Barry was not alone in thinking so: The core founders in my dataset were 26% more likely to 
go solo if they had pre-founding management experience than if they lacked it.^ 

Going solo also fit with Barry's style of work. He explained: 

I realized that throughout my career, I've been a solo guy. I puU together teams, but from a peer standpoint, I've always figured 
things out myself. With my years of experience at GTE, I could do the technology, marketing, and administrative pieces myself, so 
I never considered pulling in other founders. Also, I learned from my grandfather and dad that partnerships never work. They 
always were sole proprietors. I learned that leadership is all about taking in information and making a decision — shared 
information but not shared decisions. Make decisions yourself and live with them. Another key is speed. I want a single decision 
maker, even below me. If I have a VP of operations, he makes the call about operations. 

The decision to be a solo founder may not always be made for such efficiency reasons. Core 
founders may have such a strong preference for control that the decision to go solo doesn't even feel 
like a decision. Dr. Kevin Stone, for example, inventor of the first ready-to-drink glucosamine 
supplement and founder-CEO of Joint Juice, states enphatically, "Maniacal drive by a benevolent 



founder is a solo activity!"-^ Founders who want to retain 100% of the equity for themselves, or who 
plan to launch the startup on their own and later hire employees to fill their holes, are also more 
likely to fly solo. Figure 3.2 presents a decision chart that captures many of these considerations. 



Career factors 
Market factors 
Personal factors 
(see Chapter!) 



Missing important hunrian, 
social, or financial capitals? 



Yes 



Need those capabilities at 
time of founding? 



No 



Yes 



Prefer to work 
alone? 



No 



Pursue 
cofounder(s) 



Yes 



Become solo 
founder 



No 



Become solo founder; 
hire later 
(see Chapter 8) 

Figure 3.2. Central "Solo versus Team" Questions 



Going solo is, like most decisions we will discuss in this book, a decision with a set of short-term 
consequences and a different set of long-term consequences. In the short run, solo founding may seem 
like the easy choice, but not all solo founders with Barry Nails 's determination also have his wide 
range of skills and experience.^ By remaining solo founders, they take a much greater risk than he did 
that their startup will either fail to reach its full potential or just fail outright. For example, they may 
prioritize a need for control over growing a more valuable company, or, in their overconfidence and 
their passion for the idea, they may underestimate the need for help. 

THE ARGUMENT FOR COFOUNDERS 

While there are many valid reasons for founding a startup alone, a large percentage of founders 
decide to build a team of cofounders.* Their reasons are both tangible and intangible. The tangible 
needs that could be met by adding a cofounder include three kinds of capital. In Chapter 2, we looked 
at how each type of capital can and should affect the decision to become an entrepreneur, but similar 
considerations affect cofounder decisions: 

• Human capital. Human capital includes the explicit knowledge derived from formal education 
and the tacit skills derived from prior experience. It is a rare founder who already has all the 
skills and knowledge needed to build a new organization from nothing. For instance, founders 
with deep technical backgrounds are often unable to tackle the early tasks that require sales. 



marketing, business development, and financial knowledge. Founders who lack industry 
experience may get blindsided by problems that could have been anticipated by an industry 
veteran, cofounders with those complementary skills will maximize the startup's chance of 
anticipating or dealing with such problems. 

• Social capital. Social capital refers to the benefits derived fi*om one's place in information 
and communication networks. New startups need to reach out to recruit employees, establish 
relationships with potential partners, meet potential investors, and gain access to many other 
outside resources. Unless the core founder has a particularly rich set of relevant contacts, the 
startup will benefit fi*om cofounders' contacts and social capital. 

• Financial capital. Financial capital refers to the money or other tangible resources that can be 
used in the founding process. The founder of a capital-intensive business, or any founder who 
has not accumulated the necessary financial capital, can benefit from cofounders who have 
their own financial capital to invest in the startup. 

A founder taking this "capital-based approach" to the solo-versus-team decision needs to assess 
the human, social, and financial capital required to build the startup; compare it to the human, social, 
and financial capital he or she already possesses; see how much is missing; and then decide whether 
to build a founding team or remain a solo founder. A startup facing a very diverse set of challenges — 
for example, designing a complex medical device, developing and testing it (and securing enough 
venture capital to engage in such a lengthy and risky process), securing patents, complying with 
conplex federal regulations, creating a marketing and sales plan, navigating among tough and 
successfiil competitors, convincing hospital administrators to use the device and insurers to cover it, 
supporting the product, preparing for legal liabilities, upgrading the product, and so on — probably 
needs multiple cofounders, while a solo founder may be enough to start a medical-device distribution 
company. Industries with high levels of competition, and ones in which early entrants have an 
advantage over late entrants, may also increase the need for cofounders. These are some of the 
reasons why solo founders account for about half of the businesses in the Panel Study of 
Entrepreneurial Dynamics (PSED) small-business dataset but fewer than 20% in my dataset of 
technology and life sciences startups.^ In that latter dataset, more than one-third of the core founders 
attracted one cofounder and about 10% of founding teams even included five or more founders, as 
shown in Figure 3.1.^ 

Tim Westergren took such a capital-based approach to founding Pandora Radio. In 1999, he came 
up with the idea to construct a database of the musical attributes of songs and use it to make 
recommendations to customers (via online music retailers) based on their descriptions of what kind 
of music they liked. Tim had a political science degree from Stanford and had spent the previous 11 
years as a film-score composer and member of a rock band, which had given him substantial 
expertise in music, some sales and management acumen from managing his band and selling his 
composing services to filmmakers, a small cash cushion, and broad contacts with musicians who 
could help him catalog the music. But he had never held a "real job." Well aware that he lacked 
substantive knowledge of the complex technologies his idea required and that he had never even 
worked in a business, never mind founded one, he knew he would need help. He decided not to move 
forward until he had secured cofounders who had the human capital to fill in his technology and 
business holes, the social capital to get to potential nonmusician hires and potential investors, and 
additional financial capital. 

Even if a core founder has all the requisite skills, contacts, and seed capital, he or she may 



nevertheless decide there is too much for one person to do (or do well) and therefore seek a 
cofounder. Core founders may be tenpted to do everything themselves, but this is probably not as 
efficient as it sounds. Founding a startup is almost guaranteed to be more work than its founder 
imagines; the founder's having a full plate right from the start leaves no room for the unexpected tasks 
— whether emergencies or opportunities — that are sure to arise. In any case, working Mi-tilt all the 
time sounds heroic but is not very efficient. If crucial decisions are being made by people who are 
worn out or burned out, the startup will suffer. Jim Triandiflou recalled the early days at his startup, 
Ockham Technologies, trying to design an entire product suite for IBM with only his partner, Mike, 
and a team of consultants: "We spent Saturday and Sunday in [the consultants'] office in a conference 
room, trying to cram through what the requirements were going to be for the whole [product] suite. 
We were trying to boil the ocean. We were going to have six modules done in six months." 

At the same time, the personal preferences and psychological needs that lead core founders to 
choose cofounders can outweigh the tangible factors: 

• Task preferences — A core founder who has all the requisite skills, contacts, and seed capital 
but dislikes carrying out one or more of the critical early-stage tasks may look for a cofounder 
to take on those tasks. For example, Vivek KhuUer had worked for Bell Atlantic for several 
years, including two years as a software programmer and analyst, developing software for 
new services and systems. Vivek had the skills to create the software for his online ticketing 
idea, but he had long since moved on from programming, had held a series of management 
jobs, and was currently enrolled in an MBA program Instead of taking on the programming 
task himself, Mvek chose to add a cofounder, a prior coworker from Bell Atlantic, to develop 
the ticketing system This left Vivek free to focus on business development and finding funding 
for the startup. 

• Collaborative style — Sharing the idea with a potential co-founder can help the founder 
articulate and enhance it. When Vivek Khuller took on Saurabh Mittal, a fellow MBA student, 
as a cofounder, their discussions helped Vivek refine and change his business proposition. 
During late-night sessions from 10 p.m to 4 a.m, Vvek and Saurabh honed the Smartix 
revenue and pricing models, and developed plans for how much capital to raise and how to 
use it. Such collaboration was Vivek' s preferred working style. 

• Support and validation — ^A core founder who has a strong need for affiliation, validation, or 
psychological support may seek cofounders, even if they add little capital to the startup. One 
founder admitted he was "anxious to get somebody on board," even though there was no clear 
hole that a cofounder would fill. Another founder said, "If I can't even convince one friend to 
join me, how good could my idea be?" Brian Scudamore started out building Rubbish Boys, a 
residential rubbish-hauling company, by himself, but quickly decided to add a friend as an 
equal partner: "I didn't put much thought into it, didn't think about what holes he could fill or 
the marriage of skills. I just liked the guy and figured we'd have fun doing this together. I 
figured it might be more fun if I had someone to do this with. I needed the camaraderie, the 
support, the confidence-building of having someone else." 

It can take time for some of these psychological needs to become evident. Even a core founder who 
can handle all of the tasks may find the ups and downs of the "founder roller coaster" taking a 
psychological toll and may decide it's time to look for some cofounders. Lew Cime, the founder of 
enterprise- software startup ^^^ly Technology, worked alone for a year, developing the core 



technology for his startup's first product: "It was a critical year, and a real test of my inner fortitude. 
There were a lot of sleepless nights, from work, excitement, fear, and anxiety, wondering whether it 
had any commercial value." Instead of waiting so long to involve others. Lew could have gained 
valuable moral and emotional support to help get through those tough early days. Given the natural 
tendency for passionate founders to expect the best, if they doubt their ability to withstand the 
pressures, they should seriously consider attracting cofounders who can both provide support and 
enhance the team's capabilities. 

HOW MANY COFOUNDERS? 

Once the core founder has decided to add cofounders, he or she has to decide how many cofounders 
to add. Each new cofounder increases coordination costs and inefficiency. As highlighted by 
Stinchcombe, the larger the team, the greater the coordination costs and the higher the risk that roles 

will overlap and cause conflict within the team^ For exanple, in FeedBurner's four-person team, 
two of the founders were happy to focus on their own functions, but Dick and Steve, who shared many 
of the high-level decisions, experienced time-consuming conflict over roles and responsibilities until 
the startup grew large enough to enable Steve to focus fiill-time on business development. The core 
founder's motivation can also play a powerfiil role in the decision of whether to add each cofounder: 
A desire for control might lead the core founder to add fewer cofounders, while a desire to build the 
startup's value might lead to a bigger founding team. 

Each additional person also adds more nodes to the communication network, slows things down, 
and weakens incentives.^ Thus, each new cofounder should add important elements to the team — for 
exanple, filling in gaps in human capital or having the ability to reduce a key area of task overload 
for the other founders — that bring more value than is lost by adding the new cofounder. In the case of 
UpDown, a social-networking site for small investors, core founder Michael Reich added MBA 
classmates to his team, but their skills, experience, and networks were similar to his, which meant 
that the three-person team still needed to add a programmer, Phuc Truong, to build the product. The 
resulting stress about roles and equity splits could have been reduced if Michael had invited only 
Phuc — the one person he truly needed to realize his idea — to join the team. 

As Michael's experience at UpDown illustrates, founders tend to underestimate the costs and 
conplications added by each new cofounder; they don't realize that, as one serial entrepreneur 
observed, "Complexity increases exponentially with more cofounders." Like Michael, many founders 
also tend to overestimate the value that will be added by redundant cofounders who might not add 
new capabilities. As a result, they end up with larger teams than they should, a problem that could be 
avoided by taking a more disciplined approach to evaluating each potential cofounder 's marginal 
costs and value. In addition, core founders need to understand the potential risks and problems 
introduced by adding cofounders, which we will detail in Chapters 4 to 7. 

THE BROADER CONTEXT 

The broader environment in which the startup operates may also affect the solo-versus-team decision. 
Solo founding may fit with some environmental contexts while other environments may require a 
founding team In particular, some industries are highly competitive or strongly favor first movers. An 
industry with powerfiil network effects, for example, will provoke a "ticking clock" race to develop 



the initial product and to attract and lock in customers. In such an industry, the core founder faces a 
more urgent need for cofounders to fill in the holes in each area of capital. He or she may also have a 
greater need to fill in the psychological holes when setting out on what is likely be a very stressful 
roller-coaster ride. 

The more complex the environmental contingencies faced by the startup, the more need for 
additional founders. Teams dealing with complex environments need to process more information; 
larger teams are better able to do so. The turbulent environments in which startups operate are 
marked by rapid technological discontinuities and unstable success factors. Such environments 
increase the team's information-processing needs by creating opportunities and crises that require the 
organization to adapt its strategy and structure. Larger groups are better at solving problems because 
they can (a) absorb and recall more items of information, (b) correct more errors in inference and 
analysis, (c) consider more potential solutions, and (d) bring a broader range of perspectives to bear 
on the problem. Thus, larger founding teams, which arm the organization with additional resources, 
have higher organizational growth rates and rates of survival. 



Solo-found when . . . 

• Founder has deep human capital and 
social capital (and sufficient financial 
capital) that is relevant to the 
startup s industry 

• Founder has a strong preference for 
maintaining full control of all 
decisions 

• Founder docs not have a strong need 
for support or validation 

• The business is small and in a slow- 
moving industry 



Build a founding team when . . . 

• Founder has important holes in 
human capital, social capital, and/or 
financial capital 

• Founder prefers not to do some tasks 
required in the early days of the 
startup 

• Founder prefers a collaborative style 

• Founder has a strong desire for 
support or validation 

• The business is in a fast-moving 
industry, especially if there are first- 
mover advantages or network effects 



Figure 3.3. When to Solo-Found versus Build a Founding Team 



These differences in context further reinforce why solo founding is the exception in my dataset of 
technology and life sciences startups, but tends to be relatively common in predominantly low-tech 
small businesses. 

In summary, as shown in Figure 3.3, solo founding should be pursued by founders whose 
backgrounds, goals, and startups meet the following criteria: The founder has deep, relevant 
experience; the founder is driven to keep control of the key decisions; the industry does not demand 
fast growth; and the idea and its inplementation are relatively sinple. The opposite decision should 
be made by inexperienced founders, those who are willing to give up some control in order to attract 
an excellent cofounder, those founding startups in challenging or fast-growing industries, and those 
whose startup ideas are complex. 



SYMBOLIC FOUNDERS VERSUS HIRES 



We tend to think of "founder" as an objective status that applies to a person who was a full-time 
participant in the startup when it was founded, took part in originating or developing the idea on 
which it was based, and played a central role in getting it off the ground. This view suggests that all 
so-called founders joined the startup at about the same time. But in many instances, people considered 
to be founders joined the startup months after the core founder began it.^^ Thus, "founder" can actually 
be a very subjective — even symbolic — status. For instance, when I asked the founder-CEO of an 
email startup about the widely differing start dates and equity stakes of his cofounders, he told me, 
"To me, the 'founder' title is symbolic of their playing a key role in building a new fimction for me. 
Even if someone was employee number ten, if he was the first hire for his function, I'll call him a 
founder. Bestowing the 'founder' title is also a way for me to sweeten a job offer without having to 
give up scarce equity or cash." 

Yet, there are dangers to granting "founder" status too easily. The founder title brings higher status 
in the startup and may make the person feel more entitled to have and keep a senior position. 
However, as captured by founder and investor Jeff Bussgang's "jungle, dirt road, and highway" 
metaphor, startups go through dramatic changes, first needing people who can find their way through a 
wild jungle, then people who can create and traverse a dirt road, then people who can pave and drive 
down a highway. As we will see, startups can have major problems when early "jungle stage" 
employees later prove ineffective on dirt roads or highways and have to be sidelined or fired. When 
that person carries the founder title, these problems have even deeper, sometimes unanticipated, 
impacts on the startup's culture and employees. 

Even a team of "actual" founders can find themselves having to decide whether or not to give later 
hires the status of founder. The founding team of UpDown wrestled with just this issue. Michael got 
the idea for the startup in his first year of business school and almost immediately involved two 
classmates. Warren and Georg. Warren's interest soon waned. Georg and Michael, however, were 
still committed and knew they needed to involve a software developer to create the site. They placed 
advertisements for a chief technology officer and found Phuc, an experienced technologist. Although 
Michael and Georg could have hired Phuc as an employee, they decided to give him the title of 
founder and share the equity with him, even though he had contributed no seed equity and had not been 
involved until months after the founding. Giving Phuc the title of founder helped solidify his 
commitment at a time when Michael and Georg didn't have the resources to pay him a salary. On the 
other hand, although Warren was involved at the inception of the conpany, Michael felt strongly that 
Warren should not be deemed a founder because Warren would not be playing a central role in the 
startup, was not fully committed to it, and would not be receiving more than a sliver of equity in the 
startup. 

Empirically, we can use multiple indicators to assess which people are truly founders of their 
startups.* When all indicators for a given founder point in the same direction, we can be more 
confident that he or she was a core or "real" founder. When the indicators are mixed, as in the email 
startup above, we can be more confident that the founder title is closer to symbolic. 

CLOSING REMARKS 

Excessive optimism can blind many founders to their startup's critical needs, so they must be 
particularly vigilant about identifying the gaps in their skills, knowledge, and contacts and evaluating 
whether and when those gaps should be filled by a cofounder. Underestimating the need for a 



cofounder heightens the risk of failure. 

An immediate and critical need that cannot be filled by the founder's own human and social 
capitals does not necessarily require a cofounder. Other options may include outsourcing, finding 
advisors, or partnering with conplementary companies. But often these are not enough. 

Well-prepared, multi talented founders who do not need a co-founder to meet any of the startup's 
immediate needs may want one anyway because they prefer the camaraderie of working in a team. But 
as Brian Scudamore learned the hard way, this can be very risky. A core founder might think, I know I 
can get along without this co-founder if I need to, so what have I got to lose? Plenty, cofounders can 
slow down decision making and introduce tensions, the more so if they aren't really fiilfilling any 
essential need. Nonessential cofounders also take up equity stakes that could be used more 
productively to attract hires, investors, or other inportant players. As we will see in Chapter 6, it is 
extremely difficult and costly to undo a bad cofounder decision, as it was for Brian. Before exposing 
the startup to these hazards, a founder should be reasonably sure that a potential cofounder will add 
an important piece to the startup puzzle. 

Sometimes a founder knows that he or she is alright for now but is going to be missing a critical 
piece later on; for example, he or she has the deep technical background and skills to develop a 
prototype product but not the sales and administrative skills to attract and enroll beta-testers after the 
prototype is finished. As suggested in Figure 3.2 earlier in this chapter, such a core founder can found 
solo, develop the prototype, and then hire employees to fill the gaps as the needs (and financing) 
arise. This strategy lets the core founder keep more control over the early decisions — to stay king 
longer. 

However, the wisdom of this go-solo-and-hire-as-necessary strategy depends on the 
circumstances. When the startup's industry is growing quickly or if there are other challenges that 
place a premium on speed, even a founder who is self-sufficient at present may need a cofounder 
from the start because (a) the need for that co-founder will come soon and it may be hard to find the 
right person quickly enough and (b) a well-chosen cofounder can help build the startup more quickly 
than the core founder could do solo. 

Figure 3.4 summarizes the advantages and disadvantages of remaining solo versus building a 
founding team and suggests ways to mitigate the disadvantages. 



option 



AJt\inUges 



PutcHiLiI Wjyi /" Atifi^i;j/f DisjJtiintJges 



Gomg solo 



• Rrtain all of the equity • Have to rely exclusively on the 

• Maintain decision- founder to All gaps in human 
roakmg control capital, social capital, and 

• Avoid communication, fiiuncial capital, either slowing 
coordination, and the startup's launch while the 
incentive problems founder becomes prepared or 



• Postpone or extend decision to found; 
systematically till holes hy gaining 
relevant experience 



• Find experienced advisors and mentors 
to till holes, at least temporarily 



cxpoMng the startup to potential 
failure if the founder moves 
forward without the requisite 



• Find complementary corporate parmer» 
or outsource some tasks 



capitals or a critical competency 
• Less ability to gather and 



• As needs arise, use equity to attract 
hires and investors to fill holes (see 
Chapters 8 and ^. respectively) 



process complex information 

• Slower response rate 

• Lack of collaboration/support; 



• Found in a slower-moving, less complex 
industry 



lonely 



Building <i 
founding team 



• Fill holes in human • SacriAcc equity- 
capital, social capital, and • SacnAcc decision-making 
financial capital control 

• Increased ability to gather • Communication, coordination, 
and process information and incentive problems 

• Faster response rate 

• Gain support/ 
collaboration 

• H,ive more fun (for the 
right personality mix) 



• Carefully evaluate the marginal utility 
of each new cofounder; only add 
cofounders whose marginal bencAts are 
more than the added costs 



• Proacrively develop process for decision 
making withm the team (see Chapter 5) 



• "Try before you buy"; invest time 
in getting to know prospective team 
members and their working stylet 



Figure 3.4. Advantages and Disadvantages of Going Solo versus Building a Founding Team 

Even when bringing on cofounders is the right decision, it is almost never an easy one in the longer 
run. It immediately adds complexity to the startup and sows the seeds for a myriad of future dilemmas 
and challenges, many of them underappreciated or unanticipated when the initial decision is made. 
Understanding those challenges before making that initial decision is crucial for making the right 
founding-team decisions. In particular, there are three recurring categories of founding-team decisions 
— relationships, roles, and rewards — each involving trade-offs and tensions. These "Three Rs" are 
the subject of the rest of this part of the book. 



* Other core founders try in vain to attract cofounders, and decide either to go solo or to defer founding the startup (Ruef, 2010). 

* The indicators used in my analyses were the following: 

• Is each founder's start date the same as the startup's founding date, or are some later? 

• Was each founder at least partly responsible for the idea or intellectual property on which the startup was founded? 

• What was each founder's initial role in the startup? For example, did each founder begin with a C-level title? Were some 
founders on the board of directors from the board's inception while others were not? 

• Which people are explicitly labeled as founders of the startup? 

• What was each founder's equity stake? 



CHAPTER FOUR 

RELATIONSHIP DILEMMAS: FLOCKING 
TOGETHER AND PLAYING WITH FIRE 



Founders who decide to form a founding team must now decide whom to choose as cofoimders. 
As shown in Figure 4. 1, the prior relationships within the founding team introduce one of three major 
sets of dilemmas with which founders have to deal, and which we examine in the remainder of this 
part of the book. 

Core founders have many options for where to look for cofounders, which can be envisioned as a 
series of three concentric circles. The inner circle includes people with whom the core founder is in 
direct contact because they already have a relationship that ranges from old neighborhood friends to 
husbands and wives. The middle circle includes people met through indirect contact or indirect 
networking; that is, through a mutual acquaintance. The outer circle includes people met through an 
impersonal search process; that is, strangers who are identified for having particular traits or 
abilities — or sometimes just because the founder takes a liking to a new acquaintance. 

The case of Evan Williams illustrates both the range of options available to cofounders and some 
of the possible short-term and long-term consequences of their decisions. To found his first two 
startups, Evan chose people who were socially close to him. The first startup, a direct-marketing 
firm, was cofounded with his father and also involved his then-girlfriend, his brother, and several 
college friends. The team avoided discussing tough issues, such as how to deal with conflicting roles, 
and never developed an effective working relationship. After three years, Evan shut down the startup: 
"It was a mess with our relationships and the failed projects. ... It was a train wreck management- 
wise." His second startup, which turned into Blogger, an early pioneer in the blogging segment, was 
co-founded with Meg Hourihan, a technology consultant whom he had been dating. The problems and 
pains that Evan experienced when cofounding with his relatives, friends, and girlfriends — 
experiences that we will discuss below — led him to take a very different approach in his third 
startup, a podcasting conpany called Odeo. This time, he decided to cofound with Noah Glass, an 
acquaintance who had prior experience in the online-audio industry but whom he did not know well. 
As we will see, at Odeo he avoided some of his previous risks, but encountered others. 



founding-Team 
Dilemmas: 



Beyond-the-Team 



Dilemmas: 
• Hires? 



Should I found 
now? 



Yes 



Should I be a 
solo founder? 



No 



^ I* Relationships? 



• Roles? 

• Rewards? 



• Investors? 

• Succession? 



No 



Remain nonfounder 



Yes 



Figure 4. 1. Relationship Dilemmas, in the Context of the Broader Set of Founding Dilemmas 



In this chapter, I focus on two issues in particular: (a) whether the founding team should be 
homogeneous or diverse and (b) whether to found a startup with family or close friends, which has 
high benefits but underappreciated risks, or to found with other types of people. I examine the core 
dilemmas, describe the quantitative implications for team stability, and outline ways to manage the 
risks while getting the benefits of each option. 

FOUNDING-TEAM HOMOGENEITY VERSUS DIVERSITY 

Proverbial wisdom states that "birds of a feather flock together." Sociologists call this natural 
tendency homophily and have shown in small businesses that people of the same gender or race and 
people of similar geographic origins, educational backgrounds, and fiinctional experience are 
disproportionately likely to found companies together. Excluding spousal teams, all-male or all- 
female founding teams have been found to be five times more likely to occur than would be expected 
by chance, and teams are remarkably homogeneous with regard to skills and fiinctional backgrounds.^ 
(Also, ethnically homogeneous teams were found to be 46 times more likely, and, even afier 
controlling for the possibility that ethnic homogeneity is the result of family ties among the team, 
ethnically homogeneous teams occurred 27 times more often than would be expected by chance.) 
Thus, homophily has powerfiil effects on the homogeneity within founding teams. 

We can get a sense of how powerfijlly homophily affects the formation of founding teams by 
considering how the teams in my dataset vary in terms of work experience. One might expect, for 
example, that the greater the team's average years of work experience, the more variation there would 
be within the team due to young people with exciting ideas recruiting experienced cofounders to help 
them build the startup, or to older founders recruiting young cofounders who are more in tune with the 
latest technology, social trends, and so on. However, except for the least-experienced teams, there 
seems to be a consistent threshold — in my dataset, about a decade of difference in prior experience — 
beyond which teams resist adding people who are much more or much less experienced than they are, 
no matter how experienced the team. Whether the teams average 10 years of prior experience or 29 
years, they tend to include people who are similar in the length of their work experience. 

Short-term Benefits of Homogeneity 

Homogeneity has important benefits, perhaps the most immediate of which is speed. For the founder 
scrambling to meet the challenges of a growing startup, choosing cofounders from among people with 
whom he or she probably has important things in common is often the quickest and easiest solution.^ 

Not only does it generally take less time to find people who are like you in some important way, 
but it also generally takes less time to develop effective working relationships with people who are 
like you. When founders share a background, they share a common language that facilitates 
communication. They have higher confidence that they will be able to develop the deep level of trust 
necessary to become an effective founding team To some extent, they already understand each other 
and can skip over part of the learning curve that would absorb the energies of people with very 
different backgrounds. It is also easier to access people who are similar to you. Having a team of 
well-acquainted individuals with similar backgrounds can facilitate the early process of constructing 
an organizational identity with clear borders between members and non-members,-^ and may enable 
teams to consider alternative viewpoints without splintering."^ Increasing homogeneity may therefore 



be a particularly tempting — and, in some ways, a particularly wise — approach for novice founders 
heading into unfamiliar territory. 

Indeed, studies have found that the greater the heterogeneity among executive team members, the 
greater the risk of interpersonal and affective conflict^ and the lower the group-level integration.^ 
Heterogeneous cofounders may run into early problems if their differences result in incompatible 
working and communication styles and if they have trouble appreciating the value that the other 
person brings to the startup. Michael Reich, for example, cofounded UpDown with Georg 
Ludviksson, an MBA classmate, and Phuc Truong, a software programmer whom he had recruited 
through an ad placed on the Harvard Alumni Start-ups mailing list (a classic example of impersonal 
search). Michael took the lead in business development and financing, Georg focused on product 
management, and Phuc was the CTO whose job was to build the software to run the site. As Michael 
worked on the business plan and on securing fiinding for the startup, he became increasingly unhappy 
with the relative amount of contribution by his cofounders, and created a revised equity agreement 
that increased his own equity in the startup. This angered Phuc, who felt that his contribution would 
eventually be the most important and time-intensive part of the startup phase and was being severely 
underrated. Phuc also felt that Michael, not being a programmer, could not understand the time and 
effort it took to create a software program. Michael, on the other hand, felt that other programmers 
could do what Phuc was doing, whereas his own contributions were irreplaceable. The dichotomy in 
the two founders' backgrounds created an impasse where neither understood the other's point of 
view, sparking a crisis within the team. 

Longer-term Risks of Homogeneity 

As tempting as it is to go with the "comfortable" and "easy" decision to found with similar 
cofounders, founders maybe causing long-term problems by doing so. As we saw in Chapter 3, teams 
with a broad range of relevant fimctional skills may be able to build more valuable startups. 
Conversely, homogeneous teams tend to have overlapping human capital, making it more likely that 
the team will have redundant strengths and be missing critical skills. Although we saw above that 
heterogeneous teams were more prone to internal conflict than homogeneous teams, teams with 
homogeneous fimctional experience have been found in multiple related contexts to be less stable than 
their heterogeneous counterparts.^ To find the broad range of hard skills needed in the startup, 
founders have to fight their homophilic tendencies by taking a structured approach to defining the 
startup's needs for human, social, and financial capital, assessing the gaps, and seeking cofounders 
who can fill them, even if those cofounders are not similar in background. In addition, as explored in 
the next chapter on roles within the team, cofounders who have similar skills will also gravitate to 
similar, overlapping roles and thus tend to experience greater conflict than when there is a clear 
division of labor between them 

If the homogeneous team is also a relational one, made up of prior close ties, the problems can 
multiply. One experienced founder commented, "I have seen several startups where star salespeople 
figure they can use their experience to cofound a successfiil company. ... At first, sales take off, and 
the friendship continues. Down the road, they find out their best friends all share the same business 
skills and — ^unfortunately — the same weaknesses. Sales are great, but other areas, such as managing 
inventory or supply chain or personnel issues, necessarily suffer." 

Diversity in the fimctional backgrounds of team members is especially important in turbulent 



contexts — exactly what entrepreneurs are likely to face — ^because it enables teams to quickly adapt to 
changes.^ It also affects strategic creativity. A longitudinal study of high-tech firms from Silicon 
Valley found that cofounders coming from diverse prior companies were more likely to adopt an 
"exploration" strategy (developing an innovative product that increased variance within the 
population of organizations and generated intraindustry variety), while founding teams with shared 
work experience at a prior company were more likely to use an "exploitation" strategy (i.e., a product 
that decreases variance in outcomes, increases routinization of processes, and increases efiiciency of 
use).^ One founder said that this was consistent with her experiences: "A founding friend relationship 
can have too much stability, not enough change to grow the company or to grow with the conpany. 
The psychology of a family unit can really create an impasse in the growth of a company. Friendships 
can also suffer, though to a lesser degree, as people try to balance being nice and being business." 

To the extent that they think about diversity, founders tend to focus on diversity in skill sets. 
However, diversity in networks also deserves attention. Teams with diverse networks are often more 
creative and innovative, have better access to a range of potential investors and corporate partners, 
and are able to tap into a wider range of potential employees. Thus, core founders should seek 
cofounders who have social networks that both differ from their own networks and are not 
overweighted with one type of social contact. For example, a core founder who has deep contacts 
with Silicon Valley engineers should seek a cofounder who has few engineering contacts but lots of 
contacts in other relevant areas, such as business development, sales, and the investment community. 
Individuals with more variety in their social networks are better suited to entrepreneur ship; a study of 
founders in Germany found that people who had a relatively uniform portfolio of social contacts 

(dominated by many contacts of one type) were 14% less willing to become entrepreneurs.^^ 
Founders should consciously counter the powerful draw of homophily by watching for potential 
teammates who can add not only complementary skills but also complementary networks and strong 
relationships with a diverse set of contacts. 

The contrasts between two founding teams highlight the effects of homophily on team homogeneity. 
The team that founded Smartix, an online ticketing startup, was a group of business school classmates 
with similar backgrounds, networks, skills, and knowledge. Vivek Khuller, who conceived of 
Smartix, had graduated with distinction in electrical engineering from a university in India and had 
held engineering and management positions at two large American companies before coming to 
business school and interning at an investment bank. On the lookout for a cofounder, Vivek was 
attracted to a classmate who had graduated at the top of his class in electrical engineering at a 
university in India, had worked as an engineer at a large American company (though outside the 
United States), and had interned at an investment bank. But, by coming together on the basis of having 
so much in common, this team failed to fill a very large hole — the need for deep knowledge of the 
ticketing industry. Mvek had a potential cofounder within his grasp — an industry insider with deep 
knowledge and extensive contacts — ^but did not appreciate how important such a person's 
contributions would be and did not try to draw him into the team Only later did Vivek realize how 
his own lack of industry knowledge affected pitching the idea to stadiums: "Since I did not have any 
experience in this industry, I had no idea how these venues were organized." He found he wasn't able 
to answer basic questions about how the online ticketing would work operationally. 

In contrast, Tim Westergren of Pandora Radio did not limit himself to people with whom he had a 
lot in common. Tim understood that, while his business would have to do with music, what it had to 
be was a business; specifically, a business-to-business technology provider to the online music 



industry. He understood that he lacked the skills such a startup would need and that a merry band of 
fellow musicians would lack them, too. Therefore, he looked farther afield to find business and 
technical expertise, tapping weak ties (e.g., friends of friends) to find a serial entrepreneur to be CEO 
and an experienced technical lead to be CTO. As it happened, Tim's wife had a friend whose 
husband, Jon Kraft, had recentiy founded and sold an enterprise database startup and had experience 
with venture capitalists. After discussing Tim's idea, they decided to form a startup and recruit the 
engineering expertise they both lacked. Jon had good Silicon Valley contacts and was able to recruit 
Will Glaser, an outstanding engineer with triple degrees from Cornell whom he had met through a 
mutual contact. Jon even joked to Tim that, with Will on board, "our company just became worth $10 
million." Although forming the Pandora team took longer and required more effort than if Tim had 
simply formed a team of his musician friends, his team of relative strangers was highly diverse and 
had the key skills needed to monetize Tim's idea. If the context is not time- sensitive, taking more time 
at the beginning can sometimes yield longer-term benefits for the team 

Tangible versus Intangible Differences 

You can often tell from a resume or by asking a few questions where a person comes from and what 
he or she has done. But there are types of similarity and difference that are not so easy to pick out. A 
potential cofounder's resume may not tell you much about his or her risk tolerance, personality, time 
horizon, commitment level, and value system. Founders tend to neglect these "soft" factors because 
they are harder to assess than skill conpatibility and fimctional backgrounds, but they can become 
serious problems for even the most well-matched teams. Assessing soft factors usually requires a lot 
of time (and skill), often in the form of talking frankly to people with whom your potential cofounder 
has worked and, even more importantly, "trying before buying" — getting to know each other in actual 
work settings or preliminary projects before committing to working together. 

For example, a founding team is most likely to keep fimctioning effectively over time if the 
cofounders are personally compatible and have similar (or at least not conflicting) values and work 
styles. Founders talk about the parallels between cofounding and getting married; in either case, it is 
very hard to assess compatibility without having already spent a lot of time together. Likewise, risk 
tolerance is hard to measure, but incompatibilities in risk tolerance can be stressfiil or even fatal in 
anything as inherently risky as founding a startup. 

Differences in soft factors can build over time. Steve Wozniak and Steve Jobs, cofounders of 
Apple Computer, not only were best friends beforehand but also had complementary skills, with one 
serving as the clear technical leader and the other as the external salesperson. However, their 
partnership frayed because of differing values (most centrally, ethical steadfastness versus business 
expediency) and motivations (Wozniak's focus on technical prowess versus Jobs's focus on monetary 
rewards). Wozniak, for example, had fond memories of working in a mall as an Alice in Wonderland 
character because it was enjoyable, while Jobs thought the job was horrible because it paid so 
little. Wozniak's code of ethics was, "Extreme honesty, extreme ethics, really. ... I never lie even 
to this day," but Jobs was willing to cut ethical corners, sometimes at his cofounder's expense. This 
stark divergence in values and motivations would eventually erode the trust the two had shared in 
their friendship and early working relationship. 

In short, founders need to be aware of the decisions they are making with respect to very different 
kinds of similarity. Figure 4.2 highlights that a founding team whose members are similar in various 



objective ways (their human capital and social capital) can deprive the startup of necessary diversity. 
However, a founding team whose members are too dissimilar in terms of "soft factors" (commitment, 
opportunity costs, risk preferences, etc.) can be setting itself up for early and disruptive tensions. 
Those tensions are most likely to flare up when the team is facing decisions that could make or break 
the business (e.g., how to split the equity among themselves, which founder should be CEO, whether 
and how to raise outside capital, whether to hire a new CEO, whether to sell the business, and other 
issues examined in subsequent chapters). Ultimately, as described above, cofoundcrs with 
conplementary skills but similar value systems can form a more effective whole. However, 
homogenization and the need for diversity continually conflict, requiring founders to proactively 
assess and regularly discuss and manage the resulting tensions. 



Bc'ivfits of J Hi'iiiofieiteoiis Tcjm Risks "fj yiomoficncous Tcjm 



Tangihle Human capital Cofoundcrs with similar human capital 0\-criapping functional backgrounds increase 

factors often can communicate more quickly and the likelihood that the team u-ill be missing 

easily about issues. critical skills. 



Social capital Cofoundcrs with similar contacts may Overlapping vocial networks reduce the diversity 

have highc r conlideiKe that their mutual of infornution received b>' the team: limit its 

obligations arc more enforceable. contacts with potential customers, hires, and 

investors: and may decrease innovation. 



Intjngihle 
factors 



Decision-making 
style (e.g.. 
hierarchical vs. 
consensus) 



Cofoundcrs with similar styles will often 
make decisions more quickly and easily. 



Cofoundcrs with similar st> lcs often do not act 
as effective counterbalaiKCS toeach other's 
rutural styles. 



Risk toleraiKC (e.g., 
nsk -seeking vs. 
nsk-avcrse) 



Cofoundcrs with similar risk tolerances may 
be more stable partners. 



Cofoundcrs uith similar risk tokrances will 
not countcrbabiKC each other's tendeiKies to 
shoot from the hip or to he overhcsitant. 



Commitment level 
(commitment of 
time, capital, etc.) 



Cofoundcrs with similar levels of 
commitment will be more likely to 
appreciate each other's efforts. 



Cofoundcrs who are all moderately commiRcd 
may not be able to sustain the startup. 
Cofoundcrs who are all intensely committed 
might bum out, leaving no oik to pick up the 
slack. 



Cofoundcrs with similar values will be more 
aligned regarding their pnorities and 
preferences. 



Cofoundcrs who all have similar value systems 
may not be able to counterbalance each other. 
For example, founding teams who all believe 
in taking care of their cmpk>yccs at all costs 
might not be abk to take necessary aaions 
to prune the work force as the startup's needs 
change. 



Figure 4.2. Effects of Homogeneity within the Founding Team 



A Recurring Metaphor 

Founders of startups often reach for metaphors from family life. They refer to dating potential 
cofoundcrs before getting married. Then, of course, comes the honeymoon, during which "founders 
approach their business much like a fresh romance," as one founder put it. A founder's attachment to 
his or her startup is often compared to a parent's love for a child; founders themselves continually 
refer to their startups as "my baby." Cofounder tensions are compared to fights between spouses, 
founder agreements to prenuptial agreements, and founder breakups to divorce. 

We can learn something valuable by looking at founding teams in light of the first marriage 
described in the Bible. When God considers the ideal relationship between Adam, whom He has just 
created, and the spouse He is about to create, the phrase in the original Hebrew is Eizer K'negdo, "a 
helper against him"^^ While this seems self-contradictory, it captures a crucial aspect both of 
marriage and of cofounding a startup: the development of a cohesive working relationship that 
includes opposition and the tension that arises from partners with different skills, experiences. 



responsibilities, and motivations. Founding teams should consider whether their early decisions about 
whom to include in the team, how to structure it, and how to split the rewards will stir up 
unproductive tensions that cause the team to underperform, fail, or splinter, or whether those early 
decisions will enable the team to achieve the "helper against him" ideal that will increase their 
chances of success. 

FOUNDING-TEAM OUTCOMES: FRIENDS AND FAMILY ARE LESS STABLE 

In building their founding teams, the founders of ordinary small businesses tend to rely more on trust 
and familiarity than on functional competence and functional diversity. Likewise, in high-potential 
startups, founding with friends and family ("relational teams") is a very common — though often not 
very well-thought-out — decision. In my quantitative dataset, I found that 40.0% of the founding teams 
included at least one set of cofounders who had a prior social relationship without having shared a 
prior professional relationship (i.e., "founding with fi-iends") and 17.3% of teams included at least 
one set of cofounders who were related to each other ("founding with family"). 

Many of the same factors that lead to homogeneous teams also lead to relational teams: ease of 
access, speed of formation, and high comfort level with people who have shared experiences and 
communication styles. But in addition, cofounders with whom the founder is already close can 
provide the emotional support required to deal with the stress of beginning a startup. For instance, 
serial entrepreneur Dick Costolo says, "There are always points in the startup's life when things are 
going very, very badly, and the stress can be unbearable. Knowing the other people in the boat with 
you can be very helpful in navigating the rough waters." 

Although it is clear fi^om the data that relational founding teams are common, the hazards of close 
relational teams can outweigh their benefits and imperil the stability of the founding team Paul 
McManus, a member of a Boston venture capital firm, reflected on his experiences with founding 
teams of fi"iends: "In my opinion, those who found companies with friends will (a) lose the company, 
(b) lose their friends, or (c) lose both. I strongly advise against it and shy away fi-om deals where the 
teams are too tightly knit on the personal side. Blood [family relationships] is almost always a show- 
stopper." 

Why is founding a startup with close friends or family so risky? For one thing, tapping only the 
inner circle of potential cofounders may result in a weaker team that builds less value. One founder 
observed, "In a startup you need highly motivated people who don't need constant direction and who 
bring much more than the job requirements. In an average group of friends and associates there may 
be one in ten people who fits this description. So the chance that a candidate from this group will 
succeed is small." A venture capitalist suggested that this may be even more of an issue with 
relatives: "I would question whether the two perfect cofounders for a business could really come 
from the same family." 

A study of nearly 400 startups, conducted by my colleague Matt Marx and myself, supports 
McManus 's observation; we found that prior social relationships (friends or family) and prior 
professional relationships (prior coworkers) differed demonstrably in their inpact on a founding 
team's stability. Categorizing the founders' prior relationships along the spectrum of strong to weak 
ties — family, friends, prior coworkers, and strangers/acquaintances — ^we found that, controlling for a 
wide range of other characteristics, the type of prior relationship had a significant impact on team 
turnover. For example, each additional social relationship (i.e., a preexisting relationship not 



involving shared work experience) within the team increased the hazard rate — that is, the likelihood 
— of cofounder departure by 28.6%. Teams of prior coworkers were significantly more stable than 
both teams who had a prior social relationship and teams of strangers. It is particularly striking that 
turnover was higher for teams of friends, as those teams would be expected to put a high premium on 
team stability and harmonious relationships. Most surprising, teams with prior social relationships 
were even less stable than teams of strangers. As we will see below, it seems that a prior social 
relationship can be an obstacle to building something superficially similar but essentially very 
different — an effective professional relationship. 

This is not to say that friends and family members should never found a startup together, but that 
they should make sure that they proactively analyze the potential consequences of their prior 
relationship and take the steps described below to reduce the inherent risks. However, the picture that 
emerges from the data is that founding with coworkers can be beneficial, while founding with friends 
or family is a high-risk, "high-variance" proposition. (The complexities can multiply when a team 
includes a mix of these prior relationships.) A family/friends team could turn out to be either the best 
or the worst of both worlds. At best, the team is in perfect sync, a high-performing whole greater than 
the sum of its parts. At worst, the cofounders make suboptimal business decisions in order to protect 
their social relationships, but as the business suffers from those poor decisions, tension rises and the 
social relationships suffer, too. The best-case scenario is much more likely in teams that understand 
the "relationship risk" and proactively act to reduce that risk, as described later in this chapter.* 

Cofounding with Past Coworkers: Less Endearing but More Enduring 

The most appropriate professional structure within a startup is often at odds with the structure of the 
cofounders' established social relationships. A boss-subordinate relationship, for example, may make 
perfect organizational sense but will not suit a pair of best friends very well; nor will positions of 
equal authority suit a father and son. Unlike actual strangers, friends and relatives have to undo a 
prior relationship (at least while on the job) in order to build something very different and even 
contrary — a professional relationship. Many either choose not to do so or never even consider doing 
so. One study of high- technology startups found that the vast majority of firms in which family or 
friends of the founders were listed as key partners adopted an employee-relations model based on 
informal communication and the decentralization of authority rather than more structured and explicit 

approaches. One founder observed, "Reframing a profound friendship as a business relationship is 
difficult and sometimes painfiil." Teams of friends may start off strong and enthusiastic, but as the 
realities of life in the workplace begin to settle in, professional issues can creep into cherished 
personal relationships. It is instructive that my work with Matt Marx shows that, during the first six 
months of the startup's life, there is no statistical difference in the stability of teams with and without 
prior social relationships, but after that honeymoon period, teams with prior social relationships are 
significantly less stable. 

For all of these reasons, the transition from a previous professional relationship to a new 
professional relationship should be easier and longer lasting than the transition from a purely social 
relationship to a professional one.* Across a variety of entrepreneurial contexts, research has shown 
that teams with greater shared work experience have higher growth rates, higher levels of social 
integration within the group, and lower risk of conpany dissolution.^^ Comparing the founding teams 
of Apple Computer and Ockham Technologies highlights the very different dynamics among 



cofounders who had been friends before founding a startup and those who had afready been 
coworkers before founding a startup. Jobs and Wozniak were best friends first before becoming 
cofounders; their friendship hindered their ability to avoid conflicts in advance or at least to resolve 
them before they became unresolvable and the friendship itself became damaged beyond repair. For 
the cofounders of Ockham, the prior relationship had been not only professional but also hierarchical; 
founder-CEO Jim Triandiflou had been the supervisor of one of his cofounders. Ockham's cofounders 
were able to tackle tough issues, such as equity splits, that Apple's cofounders had felt the need to 
evade. 

The picture can be complicated by the fact that cofounders often develop very close relationships 
as their startup evolves; this happened with the Ockham team. But such "cofounders-then-friends" 
relationships are usually quite different from relationships where friends later become cofounders, 
and do not face as much business risk. Venture capitalist Tim Connors observes, "The best situations 
I've seen are founder teams who have worked together in the past, especially if that work was 
relevant to the new venture .... They're friends, but the friendships developed out of mutual 
admiration for their commitment, capabilities, and expertise. They can trust each other. Each one 
knows how the others think and work. So the whole 'cultural fit' question is taken care of right from 
the start." Business historian Richard Tedlow, paraphrasing John D. Rockefeller, puts it succinctly: 
"A friendship built on business can be glorious, while a business built on friendship can be murder." 

Where do classmates fall? In many ways, they are very similar to friends. The initial basis for their 
relationship is social, they face strong homophily pressures, and they are more likely to overlap in 
their skills than to complement each other. They may be impressed by each other's comments during a 
case discussion but not realize the gap between that and the ability to contribute constructively to 
building a startup. On the other hand, school can provide low-risk opportunities to work together on 
projects or on entries in business plan competitions, efforts through which they can gauge 
conpatibility before they have to make long-term commitments to cofounding together. For instance, 
Janet Kraus and Kathy Sherbrooke, who later cofounded corporate-concierge company Circles, met 
at Stanford Business School. They spent much of their time at Stanford coleading activities — for 
example, the student show during their first year of the MBA program and the alumni gift campaign 
during their second year — and figuring out if they were an effective team. Janet recalled another 
formative experience: 

The last thing we did upon graduation was to drive cross-country for five days and talk about what it would take for us to start a 
business together. What would be our hopes and aspirations, what kind of business would we start, what were our financial risk 
profiles, what would we be proud of, what things did we like to do, what did we not like to do, what were our greatest strengths 
and greatest weaknesses, what were our pet peeves, what we wished we could change about ourselves, what we feared, what 
would make us quit trying, what kind of people would we hire and what would we expect of them, would we consider another 
founder for skills we did not have (no), what would our values for the company be? We were really trying to make sure that we 
would be good as partners. 

Classmates who neglect such opportunities to work together, but assume that being classmates 
ensures their compatibility, face the heightened risks that Janet and Kathy were able to mitigate during 
their time as classmates. 

TAKING RELATIONSHIP RISKS VERSUS CREATING FIREWALLS 

However risky it is to found a startup with friends or family, the lure is great, and we can assume that 
the practice will remain very common. If there is no way to reduce the lure, at least it may be 



possible to reduce the risk by understanding the factors that increase or decrease it. 
The Playing- with-Fire Gap: Damage versus Avoidance 

Two factors in particular affect the stability of the founding team and depend on whether the 
cofounders are closely related (family or friends), have prior relationships as coworkers, or are 
acquaintances. Clarifying these factors will highlight the reasons why founders should try to avoid 
founding with friends or family, but also help us come up with approaches to reducing the risks of 
doing so. 

Damage If the Social Relationship Blows Up 

The closer the prior relationship, the greater the damage if tension from the business spills over into 
the relationship. There is no shortage of firsthand testimony to the prevalence and seriousness of this 
risk. As one founder who had been burned by such problems commented, "I don't mix friendship and 
business anymore. It's turned out very uncomfortable in the past. It's very hard to maintain a 
friendship after the business relationship turns sour." 

Evan ^^^lliams recalled the price he paid when disagreements over strategy and control of 
Blogger, along with its financial difficulties, caused unresolvable tension between himself and his 
cofounder, Meg Hourihan, who had been his girlfriend when the startup began. Many of his closest 
friends also worked for the startup. 

When everything fell apart, it was especially hard for me because my social life was wrapped up with my work and so I didn't 
really have a support network of friends to fall back on or find comfort in, because they were all gone with my work, with my 
employees. Things even got tense with my roommate, who was now dating Meg. My roommate didn't work for [Blogger], but his 
brother did, and his brother was actually the first guy to quit. It was just very awkward because I felt like all these people very 
close to me, in sort of this intertwined circle, were all against me, all of a sudden, as well as their extended community of friends 
who we would socialize with. Because there were more of them than there were of me, it just seemed like I was rejected from 
my entire social circle. 

Another founder elaborated, "I personally find that it is a double-edged sword. Working with 
loved ones, family, or friends can allow for better communication, and more trust — they probably 
want you to do well, as well as wanting the best for their own fiiture. But when there are conflicts, it 
is hard to leave them in the office. It becomes personal, and the other edge of the sword can cut in." 
An e?q)erienced investor reminds us of the worst-case scenario: "I've heard some successes here, but 
if you think having a friendship fail out of business failure is bad, imagine the inplications if this is 
family — ^you only have one of those." 

Avoiding the "Elephant in the Room " 

At some point, there is bound to be an inportant problem that at least some of the cofounders will 
find uncomfortable to talk about — ^the elephant in the room cofounders with prior professional 
relationships, who are used to working through such tough business issues together, are the most 
likely to bring up problems like this, followed by cofounders who were strangers and realize that, 
because they don't know each other, they have to work through such issues, cofounders with prior 
social relationships are often the least likely to deal with the elephants in the room 



Unsurprisingly, they tend to avoid difficult discussions and confrontations in order to maintain their 
relationship. They tend to worry that raising thorny issues between friends (or family) would signal 
distrust of that friend (or relative). If a business decision makes business sense but might damage the 
social relationship, founders naturally try to avoid making the business decision in the hope that the 
problem will pass. For instance, if one cofounder's ability to contribute is not growing as the startup 
grows, relatives or friends on the founding team may be very reluctant to reassign him or her to a 
different (often lower-level) role and may just wait, hoping either that the person in question will 
somehow rise to the occasion or that the occasion itself will somehow pass. Instead, avoidance often 
exacerbates the problem to the point where the startup suffers and the cofounder who was being 
sheltered is squeezed out altogether. As one cofounder put it, "I sometimes feel that if I continue to 
push to make this business work, unless my partner begins to match my energy — and soon — then both 
the business and friendship will be over. If I walk away, I may lose my once-in-a-lifetime opportunity 
to kick-start an effort that will really make a difference in transforming this planet ... as well as 
losing what has been a valued friendship." After the splintering of another team, a cofounder recalled 
a key moment when the team could have recovered: "No one wanted to be the fall guy who raises the 
issue and tries to make the team stronger." 

In such cases, concern for the social relationship interferes with making the right business 
decisions and the startup will either fail to reach its fall potential or else fail altogether. Bei Guo, a 
serial entrepreneur from China, said about cofounders who were friends, "Because they are your most 
trusted friends, you can be almost 99% open with them. But somehow there is 1% out there you think 
you will risk hurting them if you say it. It is exactly this 1% that makes all the difference." 

cofounders who have a prior social relationship also assume that "we know each other well" and 
thus don't need to discuss the "elephants." But their new professional relationship is quite different 
from their prior social relationship; in this sense, they need to think of each other as "strangers," 
which is very hard to do. The dangers may begin even before the founding, for the prospect of 
cofounding a startup with friends or family may cloud the judgment of a potential cofounder trying to 
decide whether to go ahead with it. You might join a good friend in a startup that you would never 
join otherwise, but the reasons why you wouldn't join that startup with a stranger are probably just as 
valid as in the case of your friend. Mutual affection may blind the team to what is missing from the 
idea, the business plan, or the team itself As one founder pointed out, "We trust our friends too much 
at times; founding with a stranger will force a person to critically assess both the idea behind the 
startup and what their own personal role will be. The startup will be viewed not so much as an 
exciting fiin project with a bunch of old coworkers and friends, but as a business proposal with 
significant career risks." The founder went on to explain that judgment may also be clouded in ways 
that exacerbate the downsides of homophily: "I think you're much better off founding with and hiring 
strangers; they're easier to part with, easier to be 'all business' with, and, whether you like it or not, 
will bring a lot more diversity to the company than family and friends." 

Steve Wozniak and Steve Jobs again provide a good example. For Wozniak, his friendship with 
Jobs was a critical element in favor of starting Apple Computer. "I was excited to think about us like 
that. To be two best friends starting a company. Wow. I knew right then that I'd do it. How could I 

not?"^^ But later in their working relationship, Wozniak and Jobs failed to discuss crucial issues 
about roles and rewards that were too uncomfortable for best friends to broach. For example. Jobs 
felt that his employee number ("number two" versus Wozniak's "number one") indicated that the 
company did not value his role; he wanted to be assigned "employee number zero." Wozniak, on the 
other hand, felt that Apple's later treatment of Jobs's Lisa product-development team relegated 



Wozniak and his Apple II team to second-class status. Neither Jobs nor Wozniak seemed to truly 
understand each other's contributions to the company. For Jobs, technology was secondary to the 
monetary value of the product; for Wozniak, the monetary value of a new technology was almost 
beside the point. These issues festered unresolved during their entire working relationship. In such 
teams, things are often very calm on the surface, but tensions rise underneath as tough conversations 
are avoided.^^ 

Wozniak also disagreed with Apple's reward system, particularly in the way the company awarded 
stock options to some employees but not others. But instead of coming to an agreement with Jobs 
about how to change the system, he created the "Woz Plan," selling his own shares to employees at an 
extremely low price. His relationship with Jobs prevented him fi*om acknowledging, even to himself, 
what may have been the largest elephant in the room — his increasing disappointment in Jobs's moral 
code. Wozniak recalled learning of Jobs's duplicity regarding the financial rewards from a project 
building a circuit board for an Atari game: "He got paid one amount, he told me he got paid another. 
He wasn't honest with me, and I was hurt. . . . But you know ... he was my best friend, and I feel 
extremely linked to him."^^ As they continued to avoid talking about sensitive issues, the tensions 
continued to rise between the former best friends, and they eventually parted ways. 

The Playing-with-Fire Gap 

Figure 4.3 shows how these two factors — "damage if relationship blows up" and "likelihood of 
discussing 'elephants' " — ^vary with the type of prior relationship the cofounders have. For each type 
of relationship, the greater the distance between the two factors, the more the cofounders are "playing 
with fire." The Playing-with-Fire Gap is greatest for cofounders with a prior social relationship. 
They are the least likely to deal with the elephants in the room and will suffer the most damage if 
business tensions undo their social relationships; yet their failure to deal with the elephants makes 
such business tensions all the more likely. They are, indeed, playing with fire. 



Family 
Best friends 



Acquaintances 



Past co-workers 



Damage if 
relationship 
blows up 




Likelihood of 
discussing 
"Elephants" 



Figure 4.3. The Playing-with-Fire Gap 



Past coworkers face a moderate level of damage — a higher level of damage than acquaintances — if 
the relationship blows up, but this is counterbalanced by a high likelihood of discussing the elephants 
in the room. These cofounders face little or no Playing-with-Fire Gap.* For acquaintances or 



strangers, there is relatively low damage if the relationship blows up, and also some likelihood of 
discussing the elephants in the room; because they are unfamiliar with each other, they tend to discuss 
those tough issues before agreeing to cofound or early in their cofounding relationship. 

Reducing the Gap: Forcing Sensitive Discussions and Creating Firewalls 

Some founding teams take deliberate steps to reduce the Playing-with-Fire Gap. For instance, 
Sittercity, a pioneering online service for matching parents with babysitters, was built by a young 
first-time entrepreneur, Genevieve Thiers, and her boyfriend, Dan Ratner, a serial entrepreneur. Dan 
explained what happened when he became his girlfriend's subordinate within the startup: "The 
transition was harder for Genevieve than for me. There are three parts to your life: family, friends, 
and business. Genevieve was risking two of those while I was only risking one." They tried to 
minimize the risk by actively working on both sides of the Playing-with-Fire Gap. While things were 
still calm, they agreed on a framework for speaking openly about whatever elephants were in the 
room (including the fact that they were risking their own romantic relationship by working together). 
Dan's family had worked together for at least two generations and Dan himself had worked with his 
sister in a prior startup, so he was used to tackling sensitive issues. Genevieve explained that 
"running Sittercity has gotten me used to discussing things in a calm and professional manner with 
Dan," who had become her most important coworker. She described how the duo had proactively 
crafted a "Geneva Convention" to prevent professional disagreements from becoming personal: 
"When we have a disagreement, we have to write it up and copy the entire executive team. That 
forces us to get other people involved and to stay focused on the issues at hand rather than on each 
other." This helped them tackle issues head-on rather than avoiding issues on which they might 
disagree. 

Addressing the other side of the gap, Genevieve and Dan constructed "firewalls" that would 
reduce the damage should their relationship dissolve. Initially, Dan said, "When we fought about 
something at work, it carried over into our personal life, which was draining at times. . . . Sometimes 
it was hard to figure out what level of familiarity to bring to the office; familiarity that would be fine 
at a personal level doesn't work at a business level. I'd say something to her that frankly would not 
be appropriate for me to say to her in her capacity as my boss." To avoid such problems, they erected 
firewalls that included an explicit "Disaster Plan" calling for Dan to leave the startup were there to 
be any irreconcilable disagreements. (These disaster plans for the business culminated in a prenuptial 
agreement, the ultimate in tough discussions, which covered both personal and business issues.) This 
had the effect of pushing downward the "damage if relationship blows up" line in Figure 4.3 and 
therefore decreasing the gap (and the risk). In addition, Dan delayed joining the startup full-time until 
there was a clear need for his expertise. He also understood his role to be secondary to Genevieve's 
("This is Genevieve's business, and I'm here to help her") and he worked hard to establish 
relationships with the startup's initial enployees so as to be genuinely part of the conpany and not 
just the boss's boyfriend. Over time, Dan's and Genevieve's decisions and actions reduced their 
Playing-with-Fire Gap dramatically by minimizing the distance between the two lines in Figure 4.3. 
As Genevieve explained, "Eventually, Dan integrated into the company, after which it didn't seem 
odd that he was with Sittercity full-time. We needed time to learn to trust one another in our new 
roles. After a difficult beginning, we were on our way to becoming a 'two-headed monster' where we 
would finish one another's sentences." 

Other teams have found additional ways to reduce the Playing-with-Fire Gap. For instance, a 



founder who brought in his mother to help create a marketing plan asked her to report to his cofounder 
in an attempt to "compartmentalize" their two relationships and limit the impact of work tensions on 
their family relationship. Other teams that include friends or family try to involve an outside, 
experienced party whom both cofounders respect and who forces discussions about sensitive issues, 
plays the role of "objective mediator," and can play a tie-breaking role in an impasse. There may 
well be other inportant mechanisms for managing the Playing-with-Fire Gap by crafting creative 
firewalls; researchers would do well to seek them out for study. 

CLOSING REMARKS 

Founders have to guard against many natural inclinations to follow the "easy" path. One of those is 
the tendency to build homogeneous founding teams. Doing so makes sense — one expects that people 
who are like each other will get along easily and be able to understand and trust each other — ^but there 
are fatal flaws. Teams in which the cofounders have similar backgrounds tend to lack some necessary 
skills, to have more conflict over roles (as we will see in Chapter 5) , and to have a conparably 
limited range of useftil contacts. Such teams will thus tend to build less valuable startups than do 
teams whose skills and contacts match the startup's needs rather than the cofounders' comfort zones. 
Instead of defaulting to the easy option of finding similar cofounders, core founders should 
consciously analyze and decide which capabilities they need and then actively pursue people with 
those skills and abilities. 

On the other hand — and making life more complicated — core founders who do aim for diversity in 
human capital and social capital should also aim for similarity in "softer" areas such as values, 
commitment, and risk tolerance. Of course these are harder to evaluate, but they are the areas in 
which similarity helps reduce founding-team tensions. This isn't just a matter of comfort versus 
aggravation; it's often a matter of whether the startup survives at all and, if it does, how much value it 
creates for its founders and investors. The best way to evaluate such compatibility is often to "try 
before you buy" by finding self-contained work tasks on which to collaborate before committing to be 
cofounders — a commitment that is very unpleasant to reverse. 

A natural place to look for cofounders is among family and friends. Think twice! Founding with 
family or friends involves the risk of homogeneity plus two other much higher risks. 

First, it is natural but wrong to assume that trust and relationships developed in the social realm 
will transfer easily to the professional realm. Social and professional relationships have very 
different foundations and operate in very different, often conflicting, ways. For instance, social 
relationships place a premium on fairness and equality while professional relationships place a 
premium on equity and merit, and, in many situations, equality and equity can conflict. It makes sense 
for a diverse group of siblings to divide a family inheritance evenly; it does not necessarily make 
sense for those same siblings to apportion a startup's executive responsibilities, condensation, or 
equity evenly. The mistaken assumption that friends and relatives will be able to make a smooth 
transition to a cofounding relationship can be very dangerous if it leads the cofounders to avoid 
tackling sensitive issues up front. My analyses show that friends-and- family teams are not only more 
unstable than teams of prior coworkers (which we might expect) but even more unstable than teams of 
strangers. This is because strangers usually realize that they have little or no idea how their potential 
cofounders operate in a work environment and will therefore engage in tough discussions to hammer 
out expectations before they commit to the startup. Friends and family, who already have so much 



(social) experience with each other, tend to feel there is no need for such uncomfortable 
conversations. The irony is that friends and family risk much more — a treasured relationship — than 
strangers do, yet typically take much less care to mitigate that risk. 

Second, it is extremely difficult to keep the new cofounding relationship from affecting the 
underlying social or family relationship. The extreme ups and downs of startup life can cause major 
ripple effects in social relationships, potentially setting up a vicious cycle that can sink the startup 
and strain, damage, or even destroy the original social relationship. When the best professional 
decision — say, demoting or firing an underperforming CFO — conflicts with what is best for the social 
relationship — say, that CFO is your brother — the hard decision is typically either avoided, thus 
damaging the startup, or taken, thus damaging the relationship. In many cases, both the startup and the 
relationship are dearly damaged. 

Founders who decide to mix business and social relationships despite these dangers should at least 
take proactive steps to protect themselves and the startup: 

• Compartmentalize relationships. If the team's size allows, avoid having cofounders report to 
their relative or close friend. Instead, have the cofounder report to another senior executive, 
keeping the private and professional relationships separate. 

• Envision negative scenarios. Think about the obstacles — such as family issues, medical 
crises, and legal problems — that could affect a cofounder 's ability to work on the startup. 
Resist the natural inclination of many founders to "stay positive," focusing on only desirable 
or "expected" scenarios. 

• Create a disaster plan. Put in writing a plan of action for the worst-case scenarios, such as 
an irresolvable business conflict or a breakup in the social relationship. Also, make it clear 
who has the fmal say in an impasse. Sign an exit contract that will apply if a worst-case 
scenario develops. 

• Force sensitive discussions. Make a long-term policy of being open and honest about every 
personal issue that arises in the process of working together. Increasing the likelihood of 
discussing sensitive issues — ^by establishing a mechanism for it — can reduce (though certainly 
not eliminate) the Playing-with-Fire Gap. Resist the natural inclination to put off or avoid 
discussions of difficult issues in the hope that they will solve themselves. 

• Involve a referee. In order to prevent a professional disagreement from getting personal, 
write up a memo on the issue at hand and copy the entire executive team, as Sittercity's team 
would do in accordance with its " Geneva Convention." This forces others to get involved 
and puts the enphasis on the issue itself rather than on the people advancing the issue. Or, use 
a mutual mentor or impartial advisor to mediate.* 

The data described in this chapter support venture capitalist Paul McManus's suggestion that "[t]he 
ideal [lowest- risk] founding team is a team made up of former coworkers. By working together in the 
past, these teams have developed and worked out critical professional/interpersonal relationships 
that will be critical to their success and, most importantly, they have proven they can work together 
and get things done." Even so, prior coworkers thinking about becoming cofounders should evaluate 
any disconnects between their prior professional relationships and their upcoming cofounder 
relationships. For instance, if one founder used to be the other's boss and now they are going to be 
equals, those cofounders could have a rocky adjustment period. It is particularly important that they 
agree on a code of behavior and a conflict-resolution procedure. 



Cofounding with classmates may seem much like founding with former coworkers, but it can be a 
lot more complicated. This decision can counterbalance the strengths of cofounding with coworkers 
with the dangers of cofounding with friends. If the classmates know each other only socially and have 
not really "worked" together, they may face both sides of the Playing-with-Fire Gap, for they assume 
they know each other well and thus don't need to surface sensitive issues, and they also risk blowing 
up their social relationships if things go sour within the startup. Such founders also need to "try 
before they buy" or else to erect the firewalls described above. In contrast, if the classmates have 
already collaborated — or create an opportunity to collaborate — on substantive work, such as class 
projects or leadership in campus organizations, they may bring to the startup the strengths of 
cofounding with prior coworkers. 



* It should be noted that turnover within a founding team is not always detrimental. If a cofounder did not fit to begin with, or if that 
founder is not scaling with the startup and is refusing to take a lesser role, then his or her departure may decrease tensions within the 
team. On the one hand, in a sample of small businesses in the United Kingdom, turnover was higher in founding teams that had some 
founders with prior founding experience and some without prior founding experience (and lower in teams with more similarity in prior 
founding experience), possibly because of a dysfunctional power dynamic within the team. On the other hand, studies have also shown 
that teams that have spent more time together have entrenched "standard operating procedures" that hinder creativity; the "diversity 
shock" when a homogeneous team member is replaced by someone new can help kick-start the creative process. But whether founder 
turnover is beneficial or detrimental to the team, it will almost certainly be painful As one founder observed about his experiences 
cofounding with friends compared to his experiences cofounding with acquaintances, "The highs are higher, but the lows are much 
lower." In upcoming chapters, we will see other causes of turnover, some beneficial but many detrimental 

* At the same time, attorney Bill Schnoor points out that there may be legal obstacles to founding with prior coworkers (or even hiring 
them as nonfounding employees), especially if a potential cofounder has a noncompete or nonsolicitation agreement with the prior 
employer or if the employer may have a claim to the intellectual property on which the startup is based. Before deciding to cofound (or 
hire), each founder should make sure he or she is familiar with the other's agreements and the potential problems they may pose. 

^ Although it is less common, when the prior professional relationship conflicts with the one adopted within the founding team, the 
transition is usually much rockier than Ockham's. For instance, if the former boss now reports to the former subordinate, the team has to 
work extra hard to effect a smooth transition. 

* In fact, as shown in the figure, past coworkers may even have a "negative gap" given their high likelihood of discussing elephants 
while facing only moderate potential damage if there's a blowup. 

* Venture capitalist Jeff Bussgang points out that, to encourage full candor, this impartial mediator should ideally not be an investor or 
board member. 



CHAPTER FIVE 

ROLE DILEMMAS: POSITIONS AND DECISION 
MAKING 



Fights over who should be CEO, gridlock over critical decisions, and company-endangering 
tensions: Evan ^^^lliams's e?q)eriences with two of his startups, Blogger and Odeo, illustrate the 
various ways roles and decision-making dilemmas affect startups. In Blogger 's early days, Evan 
insisted on being CEO. Cofounder Meg Hourihan eventually agreed and took the vice-president title. 
But in fact, they made decisions by consensus, took turns freelancing for Hewlett-Packard to fimd the 
startup while developing the product, and sat together on a three-person board with their first 
investor. It sounds ideal, but the results — ambiguity about who was really in charge and tension over 
how to divide the tasks that both of them wanted to lead — caused disagreements that culminated in a 
serious conflict over Blogger 's direction. Meg wanted to pursue a pragmatic focus on large 
enterprises, while Evan fought to retain his democratic vision of Blogger as a gateway to the Internet 
for anyone who wanted to publish his or her thoughts online. The company's employees, wanting a 
steady revenue stream, sided with Meg. Meanwhile, Meg argued that Evan's leadership had put 
Blogger in jeopardy and that she should now take over as CEO. While Meg and Evan argued over 
direction and who should lead, the company could not secure additional funding and eventually had to 
lay off all its employees. Unable to convince Evan to let her take over as CEO, Meg also left the 
startup, leaving Evan alone with his vision. 

When Evan finally was able to resuscitate the startup, he realized that he wanted to be "completely 
free to do what [he] wanted with Blogger." This time, he deliberately avoided adding partners, 
instead hiring employees on short-term confracts and reftising to grant equity in lieu of salary. After 
selling the company to Google for $10 million in stock, Evan invested some of that money in Odeo, a 
pod-casting startup he was founding with an acquaintance, Noah Glass. The two of them, anxious to 
avoid the ambiguity that had characterized the early days of Blogger, worked out clear roles and 
titles. Noah was to be a fiiU-time CEO, while Evan would act as an advisor. But when Odeo began to 
take off, Evan committed to fiill-time involvement, sparking another struggle over who would be 
CEO. 

Evan is hardly the only serial enfrepreneur to find himself wrestling, from one startup to the next, 
with the role arrangements, title assignments, and decision-making approaches within a founding 
team. Of all the dilemmas we will cover in this book, role dilemmas show the widest variety of 
approaches, themselves affected by the widest variety of factors. There are so many different ways 
for cofounders' roles to go right or wrong — and so many different reasons why they do. Evan's own 
experiences touch opposite ends of a spectrum At Blogger, he and Meg had different titles but equal 
decision-making power. As a result, decisions that required someone to have the final say led to a 
conflict that blew up the founding team During the early days of Odeo, Noah was clearly the lead 
decision maker. However, as we will see below, even this arrangement can cause problems as a 



startup evolves. 

In this chapter, we will look in detail at the dilemmas faced by founders when structuring proper 
roles for themselves. Once again, we will focus on what founders can do (their range of options), 
what they actually do (how often each option is chosen), and what they should do (the inplications of 
their decisions). Multiple metrics will help us see how the different approaches are used and how 
frequently. These metrics include the official executive titles adopted within the founding teams at the 
time of founding and the extent to which the team's decision-making structure is hierarchical or 
egalitarian. I build on the existing literature on role dilemmas within teams while using my own data 
and case studies to show how each approach affects the founding teams that adopt it. 

EXECUTIVE TITLES: WHO WANTS THEM? 

How much do titles matter? Founders often insist, "I don't care about my official title." But at 
Blogger and Odeo, the jockeying between Evan and his cofounders for the titie of CEO — not at all 
unusual — highlights how seriously many founders do take their executive tities. In fact, they should. 
Tities are imbued with a symbolic significance and can translate into real authority. ^ (They also speak 
volumes to outsiders, who often interpret them quite differently than the cofounders themselves do.) 
Which cofounder receives which title — especially the CEO title — is usually one of the toughest early 
negotiations between cofounders, and rightly so because it is a very important one. 

Some founders shun the CEO titie. For instance, a scientific founder may aspire to be chief 
scientific officer (CSO) or a technically oriented founder to be chief technology officer (CTO) rather 
than CEO. However, founders commonly complain, "Everyone wants to be CEO." One founder 
admitted, "I have noticed that we are all trying to jockey for the leader role." After the negotiation, 
non-CEO founders often complain, "I feel like the lesser of the founders." Frank Addante recalls that 
when he started Zondigo, "I wanted to start something on my own; this time, be 'The Guy' instead of 
'The Guy Behind The Guy' Prove to myself that I can take on a challenge like that and lead the 
charge, that I could do it on my own. I thought, 'I'm going to lead the army and go after that market 
and conquer it.' It was personal motivation more than anything else." Had he not received the CEO 
titie, Frank would have been far less motivated to build Zondigo and may even have even decided not 
to cofound it at all. As we will see below, "idea people" frequently grab the CEO title. 

Because founders usually want senior titles, startups tend to create C-level positions first, whether 
or not these executives have anyone reporting to them, any products or services to sell, any customers 
or clients to serve, or any budgets to spend. In my sample of technology and life sciences startups, 
89% of the multifounder teams had at least one founder with a C-level title; 79% had a CEO, 45% 
had a CTO or CSO, 14% had a chief operating officer (COO), and 8%) had a chief financial officer 
(CFO). Of the teams, 18%) also had a non-CEO chairman. 



10% 



0% 



2-founder teams 



3-founder teams 



■ Top-heavy (C-level titles > Non 
C-level) 



68% 



73% 



□ Even (C-level titles = Non-C-level) 



15% 



9% 



□ Pyramidal (C-level titles < Non- 
C- level) 



17% 



17% 



Figure 5. 1. Degree of Top-Heaviness in Two-Founder and Three-Founder Teams, at Time of Founding 

Larger, mature companies almost always have a "pyramidal" structure with a single decision 
maker at the top, a layer of people reporting to that chief executive, and a larger group at each 
subsequent level. ^ In contrast, 74% of the startups in my sample had adopted "top-heavy" or "even" 
structures at the time of founding; 67% had more C-level founders than sub-C-level founders, and 1% 
had as many C-level as sub-C-level founders. Figure 5.1 shows that this top-heaviness existed in both 
two-founder and three-founder startups. Even more extreme, in 36% of the startups in my sample, all 
founders received a C-level title, possibly as a way to resolve the jockeying over who would be 
CEO. This was particularly pronounced for two-founder teams, 60%) of which consisted of two C- 
level officers. 

As we will see in Chapter 8, "Hiring Dilemmas," such early "title inflation" can come back to 
haunt teams who have to replace or demote those founders or early employees who were given senior 
titles but are having trouble scaling with the startup's growth and challenges. Startups with a top- 
heavy structure are most at risk for this particular heartache. 

A founding team's top-heaviness is, in itself, an indication of the founders' high level of confidence 
in themselves and their startup, which can, in turn, render them less likely to consider that such 
problems will occur and less prepared to deal with them when they do. Thus, such solutions to the 
immediate problem of "who's going to call the shots" can have severe long-term implications for the 
conpany. 

EXECUTIVE TITLES: WHO GETS THEM? 

How do teams assign their top positions? My quantitative analyses suggest that three major factors 
are (a) each founder's level of commitment, (b) which founders are the idea people who had the 
original idea or developed the intellectual property on which the startup was founded, and (c) each 



founder's human, social, and financial capital. 

Founders who quit their jobs or academic pursuits to make a full-time commitment to the startup 
are more likely to receive senior titles. In turn, a founder's title may affect his or her level of 
commitment; a founder who is unhappy with his or her initial position may be less motivated to fully 
commit to the startup. At Ockham Technologies, Ken Burows ended up with the COO title, while his 
cofounder, Jim Triandiflou, took the CEO title, primarily because Jim excelled at sales and Ken was 
more of a nuts-and-bolts person. While Jim's commitment to the startup grew with each milestone 
they passed. Ken became less committed (and also began devoting more attention to family life). Ken 
finally opted out conpletely when Jim began urging him to quit his job in order to work full-time on 
the startup. 

Regarding the inpact of idea people and the founders' human, social, and financial capital, 
economist Thomas Hellmann and I analyzed my founding-titles data to see which factors predicted 
which founder would receive each of the three most common founding titles: CEO, CTO, and 
chairman. In addition to whether each founder was an idea person (of which there could be more than 
one on a team), we focused on three other characteristics that indicated whether or not the founder 
was bringing valuable human, social, and financial capital to the startup: whether the founder had 
founded a prior startup (i.e., was a serial entrepreneur), the founder's years of prior work experience 
(in any context), and the amount of seed capital that the founder had invested in the startup. We found 
that all four factors had an impact on founding titles, but in different ways. 

Most centrally, our regressions showed that idea people were far more likely than non-idea people 
to become CEOs. As shown in Figure 5.2, 47% of the idea founders but only 12% of the non-idea 
founders became CEOs.^ Many idea people assume that they deserve to be CEO, regardless of 
whether they are true CEO material. One said, "Of course the startup idea's creator will act as a CEO 
and have veto on everything and approve the other's visions." Another idea person argued 
vehemently, "When you take an idea and then spend 18 months with no salary, living and breathing the 
idea, raising $150,000 from friends and family, putting together an executive team, etc., you have one 
driven CEO. No one is going to care about the company more than the idea guy, no one will be driven 
more than the idea guy, and no one will be able to make the tough decisions and shape the direction of 
the company like the idea guy. When you go through what most idea guy/entrepreneurs go through to 
make their idea a success, their idea becomes like a child to them and no one will love their child 
more than a parent." 

Idea people often believe that they have the best market knowledge and insight on the team, that 
they should be shaping the startup's culture, that they will be the best people to shape and realize their 
vision, and that they can best do so from the CEO position. (In addition, as we will see in Chapter 6, 
idea people and founder-CEOs often get a larger equity stake than other founders.) Even non-idea 
founders can see why idea people would make the best CEOs: "I feel the successful assignment of the 
CEO, especially in the early stages, also needs to take into account who is the most passionate about 
the venture. In most cases, that passionate person is the idea person. I have seen quite a few teams 
willing to sacrifice some of the mainstream skills of a CEO for the passion and drive that an 'idea 
person' can bring to the table." All the same, as we will examine in depth in Chapter 10, "Founder- 
CEO Succession," the idea person's fit with the CEO position, as good as it might be in the early 
stages of the startup, often weakens considerably as the startup matures, introducing its own difficult 
transition issues. Inertial tendencies and a founder-CEO's natural resistance to giving up the CEO 
title can make matters even worse. 



50% 











1^ 


fI 





Chair 


CEO 


CFO 


CTO 


COO 


VP 


Other 


□ Norvidea people 


2% 


12% 


4% 


16% 


7% 


23% 


36% 


■ Idea people 


7% 


47% 


1% 


15% 


5% 


9% 


16% 



Figure 5.2. Percentages of "Idea" and "Non-idea" Founders Who Received Formal Titles at Time of Founding 



Being the idea person doesn't only increase the chances of becoming CEO; it also can help buffer a 
founder from receiving a much lower-level title within the organization. As shown by the two 
rightmost sets of bars in Figure 5.2 (and verified by regression analyses), idea people were 
significantly less likely to receive a VP-level title or a lower title, with 59% of non-idea people 
receiving such titles but only 25% of idea people receiving them. 

Looking at the other background variables, Hellmann and I found that, even after controlling for 
who had the idea and other factors, prior founding experience had a significant positive impact on 
who would be named CEO and who would be named chairman, but had no inpact on who would be 
named CTO. The years of prior work experience had a significant positive inpact on who would be 
named chairman, but not on who would be named CEO or CTO. Finally, founders who invested more 
seed capital were more likely to be named CEO or chairman; CTOs were more likely to have 
invested less capital than had other founders. 



OVERLAPPING ROLES VERSUS DIVISION OF LABOR 



Title assignment is much easier when the founders have very different skills. The Apple team of Jobs 
and Wozniak is a classic example: Wozniak was the technological guru who invented the early 
personal computer, while Jobs was the marketing visionary who understood how to monetize 
Wozniak's invention. Jobs and Wozniak had perfectly conplementary skills and ambitions — Jobs 
wanted to build the company, while Wozniak wanted to build the product — leading to a natural 
division of labor. In contrast, when founders have very similar backgrounds or are flexible jacks-of- 
all-trades, they tend to adopt overlapping day-to-day tasks. During the early stages of a startup, when 



there are too many things to do and not enough people and time to do them, when the startup is cash- 
poor, and when the strategy and business model may have to turn on a dime, having an organization 
with flexible and overlapping roles — changeable as needed — can be a big advantage. Figure 5.3 lists 
some of the strengths and weaknesses of adopting overlapping roles versus a division of labor. But, 
as we will see below, this advantage can become a liability. 





Strengths 


Weaknesses 


Overlappintr 


• Offers flexibility 


• Diffused responsibilifv may 


roles 


appropriate to early- 


weaken incentives 




stage startups 


• Overloaded startups should be 




• Team members can 


trying to minimize redundant 




pitch in wherever 


responsibilities 




needed 


• May increase tension as founders 




• Taps collective 


step on each other s toes 




knowledge of all team 


• As the startup evolves and 




members 


Ixrcomes more differentiated, 
team members may resist having 
to focus on specific functions or 
areas, also increasing tension 


Division of 


• Enables assignment 


• May be hard to get individual 


labor 


of titles, tasks, and 


functions to collaborate on cross- 




responsibilities 


cutting tasks 




• Provides better 


• In homogeneous teams, may 




accountabilit)- 


cause early, suboptinial role 




• In bcterogencous 


assignments 




teams, enables the team 


• Failure to evolve can lead 




to tit role assignments 


to disconnects betw een 




to founder strengths 


organizational structure and task 
demands 



Figure 5.3. Strengths and Weaknesses of Overlapping Roles versus Division of Labor 

Two contrasting cases offer insights into the benefits and drawbacks of teams of people with 
general skills and overlapping job tasks, as compared to teams of specialists with a clearly defined 
division of labor. 

The Smartix team adopted an overlapping approach. All three cofounders were highly educated 
engineers, had work experience in top-flight organizations, and were MBA students at the time. Being 
so similar in background and skills, they gravitated toward overlapping roles in the startup. (They had 
brought in a fourth, "silenf ' partner to develop the software for the product, leaving the three 
cofounders to work on the business plan and fund-raising.) They worked together without clear 
separation of tasks or roles, although Vivek Khuller, the founder who had had the original idea and 
was financing the startup, acted as the de facto leader. 

Because Vivek gave himself only a year to test his smart-ticketing concept (due to the family 
pressures discussed in Chapter 2), the Smartix team did not last long enough for this arrangement to 
become a problem. However, overlapping teams can run into trouble as the startup evolves and 
becomes more differentiated. One founder in such a team described the situation: "Two of my college 



friends [and I] are invested into a business we came up with. As it stands right now, my personal 
feeling is that it will go sour soon. Dividing workloads is tough, especially when we all have the 
same talents. The ability to stay open-minded about each other's work and what will be best for the 
client has been a frustrating one. A lot of butting heads as to 'which one is better' or 'I feel mine is 
better than yours.' This last one isn't openly said but we all know it is there by our critiques. I feel 
that this business would be better served if we each had our own strong talents. Because we would 
all know what needs to be done and we would leave the other jobs for the person to conplete." 
Another founder described how his founding team had evolved overlapping roles: "When all three 
founders cannot meet with a client, the person who does meet the client takes responsibility for 
fulfilling the client's needs." For that team, however, flexibility came with a cost: "The other two 
don't shoulder the responsibility as much." 

Closer to the other extreme, the Pandora Radio team adopted a clearly defined division of labor. 
Jon Kraft had experience developing startups and had connections with Silicon Valley engineers and 
VCs, Will Glaser had extensive skill and experience as a software engineer, and Tim Westergren had 
a decade of experience in the music industry and a broad network of musicians and recording 
engineers. The founders could quickly develop an unambiguous delineation of roles and 
responsibilities: Jon, the "business guy," would be responsible for managing finances, business 
development, and adrmnistrative details; Tim, the "idea guy," would be in charge of crafting and 
articulating the vision of the music tool's capabilities; Will, the "engineering guy," would be in 
charge of building the prototype. Within his domain, each founder had almost absolute power over 
decision making and conplete control over hiring. In this model, the goal is minimal overlap. As one 
founder described her startup's founding team, "Each person is a dictator in their own task, outlined 
via 'contract' beforehand. If you're in charge of it, you get final say." 

With a clear division of labor comes clear accountability: Everyone (initially the cofounders, later 
the board of directors) can see who is responsible for each success and each failure, at least for those 
tasks or milestones that clearly fit into a particular domain. At Pandora Radio, each technical success 
or missed deadline could be traced to Will and each new partnership agreement or administrative 
mistake could be traced to Jon. When there is a breakdown within one of his tasks, each "dictator in 
his own tasl^' finds it harder to blame others. 

The experiences of the FeedBumer team highlight both a hybrid possibility and the dynamic 
characteristics of the division of labor. FeedBumer 's founding team was made up of two specialists 
and two generalists. Dick Costolo was a conputer science major who had worked on systems 
integration and development projects at Andersen Consulting, where he had managed large teams and 
worked directly with clients, and where he had met his future cofounders. At Andersen, Matt Shobe 
was working as an interface designer, and Eric Lunt and Steve Olechowski were both recent college 
graduates working on software architecture. As Dick e?q)lained, "We had very conplementary 
conpetencies. For exanple, Eric was a solid architect, but didn't like to do user-interface work 
[which Matt did so well]." When they cofounded a startup, neither Matt nor Eric wanted to pursue the 
"business stuff' that interested Dick and Steve. Consequently, they adopted a mixed structure, where 
the two specialist founders (Matt and Eric) focused on specific roles while the other two (Dick and 
Steve) moved across the remaining tasks. As Dick described it, "Steve and I can hop around a lot and 
do a lot of different things well — ^we're what you might call 'best athletes.' Steve and I could switch 
roles — he could run product management and I could run business development — and we'd be just as 
good. Eric and Matt are a different story. Eric is a software architect through and through; he couldn't 
manage a project any more than he could build a spaceship and fly to the moon. He has no interest in 



dealing with people and managing them. Matt is a user-interface designer, period." 

Dick and Steve greatly enhanced the flexibility and responsiveness of their startup by switching 
roles as needed, but their overlapping skills and ambitions caused problems. Dick explained how the 
roles within the team evolved over time: "During the proto-stages, some of our roles just fell into a 
natural form: Eric was the lead software architect and de facto CTO and I played the role of fimd- 
raising and running the company. But Steve wanted to have a more senior role, so Steve and I began 
sharing some of my tasks and we had to figure out who was doing what. ... It was contentious at 
times. I would think I should be doing something, and he thought he should. So I just said, 'Any 
product work related to a partner, you take it, Steve.' It was very organic, a result of contentious 
discussions rather than our consciously mapping out territory." Finally, Steve's role evolved into 
business development. As Costolo said, "His role took two years to evolve, but it gave Steve a 
significant piece of the company to run, and it allowed me to focus on product and operations." Thus, 
over time, the FeedBurner team developed a more effective and lower- tension division of labor. 

This is not to say that tense discussions are to be avoided within founding teams. Discussions that 
are contentious because of personal fi^iction (also known as "affective conflict") tend to signal a lack 
of trust and can cause destructive tension and finstration within the team.^ Conflicts about ideas or 
about tasks ("cognitive conflicf '), however, can often improve the team's decision making, intensify 
and enhance the relationships, and increase satisfaction. Specifically for the nonroutine tasks on 
which founders have to collaborate, such as setting strategy and resolving management issues, 
cognitive conflict is positively related to performance. 

However, a startup's evolution can cause problems for even the most elegant early division of 
labor, if that division of labor fails to evolve. For instance. Pandora's task-oriented model, which 
seemed unassailable, failed to keep pace with the needs of the growing business. The division of 
labor became very rigid, which became problematic when the startup needed to adapt its strategy. 
Resolving issues and decisions that crossed the domains of two of the founders — such as fund-raising 
— introduced added tension, slowed down decision making, and distracted them from big-picture 
issues. In addition, as Tim's leadership skills developed and his ability to fund-raise grew, his role 
increasingly overlapped with Jon's. Eventually this situation added to pressures on Jon that caused 
him to leave the startup. 

DECISION-MAKING APPROACHES: EGALITARIAN VERSUS fflERARCHICAL 

Founding teams have to balance a wide variety of trade-offs when deciding /?(9>v to make major 
decisions, balancing dynamism in decision making with clarity of direction and balancing collective 
decision making with accountability and efficiency Even if the team is not thinking about such trade- 
offs, it is still making them and will face the consequences. The decisions teams make are often ad 
hoc but sometimes derive from the founders' backgrounds or from their official roles within the 
startup (a common unintended consequence of how titles are assigned to the cofounders in the first 
place). 

To bring some order to this variety, I focus first on two contrasting ideal types, then consider some 
mixed cases. ^ The first ideal type is the egalitarian or consensus approach. The members of an 
egalitarian team ignore their official titles (in rare cases, they don't even give themselves titles), 
make decisions collectively by coming to a complete consensus, and act as peers rather than 
superiors and subordinates. The second ideal type is the hierarchical or autocratic approach. 



Hierarchical teams have a formal process for making decisions and a clear hierarchy, with a single 
person responsible for final decisions.^ These opposing approaches each have important strengths 
and weaknesses. Each can work well if it is aligned with the founders' abilities and preferences and 
with the startup's needs. Even so, as the startup's needs evolve, the decision-making approach also 
needs to evolve, and early decisions — even decisions that were good at the time — often delay or 
derail subsequent adjustments on which the startup's success depends. 

The Benefits of EgaKtarianism 

Some teams prioritize team cohesion and early trust building and believe that these will be facilitated 
by a collective decision-making approach. Other teams see collective decision making as the best 
way to avoid mistakes. In the dynamic, high- velocity environments faced by high-tech entrepreneurs, 
dominant, autocratic CEOs are likely to be less successfiil than their less dominant counterparts. In a 
study of 26 large firms in the innovation- intensive conputer industry, CEO dominance was associated 
with a 19% decrease in firm performance.^ Similarly, a study of eight firms in the microcomputer 
industry found that greater centralization of power, in which the CEO makes key strategic decisions, 
was associated with the formation of competing political factions within the top management team, as 
well as with poor firm performance.^ 

Perhaps in turbulent environments such as high tech, the complexity of the information cannot be 
processed effectively by even the most talented autocratic CEO, but instead should be handled by a 
broader team of decision makers.^ An experienced VC observed, "While dictators can move 
resources quickly behind an idea, they can also move resources quickly behind bad ideas and can 
ignore feedback." Having multiple decision makers can help avoid this problem, acting as a check on 
the so-called dictator. One founder highlighted the trade-off between decision quality and the speed 
of decision making by saying, "We had a very good experience as a triumvirate of founders. The mix 
of points of view probably slowed down our decision- making process but we felt that the decisions 
that emerged fi*omthe collaborative approach were worth the added overhead." 

When Are Founders Likely to Choose Egalitarianism? 

Egalitarian approaches are particularly likely to emerge in two situations: when the founders have 
similar backgrounds and when the team does not name an official CEO. It seems fairly natural that 
founders with similar backgrounds would be more likely to assume overlapping roles and treat each 
other as equal decision makers, but in startups that choose not to choose a CEO, egalitarianism may 
be a consequence rather than a cause of that choice. Recall that 21% of the teams in our sample 
avoided naming a CEO when they first assigned themselves titles. In some cases, this may not reflect 
a commitment to egalitarianism but rather may have been a way to sidestep divisive negotiation over 
titles, which might occur when multiple cofounders are very control-motivated. 

If so, this may show either weakness or wisdom On the one hand, not choosing a CEO might 
indicate that the founding team has trouble resolving tough issues, of which the "CEO" negotiation 
will be only one of many. Avoiding naming a CEO, as we have seen, also tends to compromise the 
clarity of the decision-making structure. On the other hand, venture capitalist Tim Connors points out 
that premature assignment of a CEO — as might have been the case at Odeo with the early crowning of 
Noah Glass as CEO — can also be problematic. This is particularly true if none of the founders 



legitimately has the skills to be CEO: "If there isn't an obvious quarterback [who can call plays for 
the team], it is dangerous to choose one of the founders as the quarterback. If there isn't a capable 
quarterback on the founding team, it should tell you that there is a hole in the founding team. I had a 
friend who founded a company with three other engineers and they arbitrarily declared one of the 
engineers as the quarterback. That founder had no experience managing a team and he started making 
bad decisions. That team would have been better off making collective decisions until they found the 
right quarterback, either as another cofounder or as a hired CEO." 

Another way of not choosing a CEO is to choose two. An official co-CEO arrangement is rare* — 
only 1% of the startups in my dataset had founders sharing the CEO title — and if a team is using such 
an arrangement to avoid a tough negotiation over tities, it can be a signal of fiiture trouble. Two kings 
fighting over a throne can leave the kingdom devastated. But if complementary cofounders are indeed 
operating as joint chief executives, and if they arrived at the arrangement after frank discussions about 
the potential pitfalls and mechanisms for avoiding those pitfalls, it can be an effective arrangement. 
For instance, at online entertainment pioneer iWon, founders Bill Daugherty and Jonas Steinman held 
the "co-CEO" titles from founding to exit, leveraging their divergent perspectives to achieve better 
decisions than they could have achieved on their own. At the time, Daugherty refiected, "We've had 
very few issues where we've had differences. Where we did, the one who's been most passionate 
about his position has carried the day."* 

Founders may also leave the CEO position vacant to signal their openness to bringing in an 
experienced CEO. But for whatever reason the CEO position is left vacant, that choice leaves the 
decision making ambiguous and often encourages an egalitarian approach within the team. (As 
described in Chapter 7, teams that split the equity equally may also be more likely to adopt an 
egalitarian approach.) 

The Problems of Egalitarianism 

Building consensus takes time and carefiil deliberation, which may help decision makers take into 
account the various, far-reaching consequences of important strategic decisions. But founders of high- 
potential startups often have no time to waste considering all the possible courses of action, as 
fieeting opportunities for growth must be seized to keep cash-poor startups afloat. Professor Kathleen 
Eisenhardt found that consensus-based teams make decisions too slowly for startups operating in 
"high-velocity" environments, those in which there is such rapid and discontinuous change (in 
demand, competitors, technology, or regulation) that information is often inaccurate, unavailable, or 
obsolete. Other researchers reached a similar conclusion: "Consensus is a worthy goal, but as a 
decision-making standard, it can be an obstacle to action or a recipe for lowest-common-denominator 

compromise."^ ^ 

At Pandora Radio, although each of the three founders had almost absolute power over decision 
making within his domain, the team required a unanimous consensus to make decisions affecting the 
company as a whole. Even a 2-1 split on an issue would result in additional debate to convince the 
holdout. Tim stated, "If one person disagreed, we had to make sure we weren't missing something, to 
give that person's opinion a fiill airing." The founder of another startup warned against this sort of 
decision making: "I'd rather have a decisive dictator who makes the occasional mistake (and learns 
from it) than an indecisive committee of founders that can't make a timely decision. Even a lame-duck 
CEO is better than [egalitarian] management. . . . [Egalitarian] is inherently less decisive because 



every move must be discussed and analyzed until a consensus is reached. A dictator has the power to 
make any decision without discussion or consensus building. Startup management must be decisive." 
In a panel discussion among four serial entrepreneurs, the one area of consensus was that "a startup 
has to be a dictatorship." One participant said, "Either I have to be making the final decisions or I 
have to be working under my cofounder who is. 'Co-CEOs' just doesn't work."^^ 

This seems to be a lesson many entrepreneurs have to learn for themselves.* Although most veteran 
entrepreneurs warn against egalitarian approaches, most founding teams choose them during the early 

stages of their startups. 



Egalitarian < 

Advantages: 

• Can help build trust among groups of 
strangers 

• For teams of friends, affirms expectations of 
equal treatment 

Disadvantages: 

• Consensus-building often takes too much 
time; this may be particularly problematic for 
high-velocity entrepreneurial environments. 

• Accountability is less clear 



Hierarchical 



Advantages: 

• Decision-makers can quickly mobilize 
resources behind a new initiative. 

• Clear accountability 
Disadvantage: 

• Complex environments cannot be processed 
by one person; input from multiple people 
with specialized knowledge usually leads to 
better decisions. 



Figure 5.4. Egalitarian versus Hierarchical Decision Making 



Figure 5.4 summarizes the advantages and disadvantages of egalitarian versus hierarchical 
approaches to decision making. 

Rocky Transitions from Egalitarianism to Hierarchy 

As we have seen from the experience of these founders, the startup's stage of development has a 
powerful impact on the effectiveness of egalitarian versus hierarchical models. Even when teams 
begin with a clear and effective egalitarian structure, they usually have to adopt a more hierarchical 

structure later on, which is often a painfiil and difficult process. Failure to do so, however, can be 
even more painfiil and costly. 

The two founders of Ockham Technologies began with a formal reporting relationship: Mike, the 
director of product management, reported to Jim, the CEO. In practice, however, they developed a 
"co-CEO" decision-making structure. Prior to landing their first customer, Mike and Jim met with the 
soflware developers, wrote product specifications, came up with a blueprint for how to manage the 
company, and toured the country together to sell the idea, even sharing a hotel room. Once they signed 
a million-dollar contract with IBM, however, Mike's job was supposed to narrow to managing the 
product development, while Jim would run the company, sell the product, and secure fimding. Jim felt 
that the close collaboration he and Mike had at the beginning seriously hindered their ability to make 
this transition. As Jim explained, "We've gone through our first growth spurt and I think we need to 
divide and conquer. We need to become a real company. We need to split things up instead of both 
being involved in everything. That means that Mike can't be involved in all decisions." However, 
Mike's prior involvement in key decisions hindered his ability and willingness to reduce his 
involvement in those decisions. 



At another startup, it took the company's imminent demise for the team to be willing to name an 
official CEO and transition to a hierarchy. The team of three cofounders had been making decisions 
collectively since founding. They had been pushed several times by advisors to inprove decision 
making, efficiency, and accountability by moving to a more formal and hierarchical structure, but had 
resisted all such changes. Only when the investor-led board threatened not to fund the startup's 
upcoming round of financing — in essence, to shut it down — did the founding team name a CEO and 
relegate the other cofounders to secondary roles and then to nonexecutive consulting roles. 

Indeed, although only 79% of the teams in my dataset began their startups having named a CEO, 
94% had named a CEO by the time they had raised their first rounds of financing, a change that 
presumably brought more clarity to their decision making. (Such "professionalization" is often one of 

the major changes driven by outside investors.) 

Expanding on these dynamics, the founder of another startup said, "My chief regret is that we didn't 
adjust our management structure and style quickly enough to adapt to the changing situation. My view 
now is that, as an early startup where the imperative is avoiding making the big mistake, having a 
group of [equal] partners with divergent views is a good thing. At that point, you're all very polite 
and respectful anyway. When the time comes that your business is working and you need to act fast 
and make decisions quickly to grow and seize opportunities, the best thing is to have a dictatorship. 
Making the transition at the right time is the elusive challenge here." 

Balancing the Best of EgaKtarianism and Hierarchy 

While it may be the path of wisdom for founding teams to evolve from egalitarianism to hierarchy, 
there is always the risk of going too far in that direction. We have already mentioned that, in a high- 
velocity environment, an overly autocratic CEO runs the risk of being unable to process the flood of 
information well enough on his or her own and the risk that his or her leadership will provoke 
political factions within the executive team. Kathleen Eisenhardt, also noting that autocratic CEOs 
may fail to achieve buy-in for key decisions, describes a two-step "consensus-with-qualifications" 
approach, balancing the efficiency of an autocratic approach with the buy-in of an egalitarian 
approach. First, the team attempts to reach a full consensus. If consensus is not forthcoming, the CEO 
and relevant functional VP make their decision, but only after getting input from everybody in a public 

forum. Describing a different model for balancing the risks and benefits of each approach, venture 
capitalist Tim Connors says, "It is important for someone to be the 'quarterback,' as you need one 
person calling plays on the field. A quarterback needs to call a play, you go run the play, then 
everyone comes back to the huddle and shares info, and another play is called. If the quarterback isn't 
listening to the inputs and adjusting the play calling, the team loses and a different quarterback is 
selected." 

DUOS VERSUS TRIOS 

If a team adopts a model with multiple decision makers of equal power (whatever their tifles), how 
many should there be? There are important differences between groups of two founders and groups of 
three. In general, larger teams have more cognitive conflict; that is, different people know or believe 
different things and there are more views the group needs to at least keep track of, if not reconcile. 
This naturally makes group cohesion and communication harder, which in turn invites more 



interpersonal misunderstanding and conflict. ' In addition, a conplementary duo in which each 
cofounder has a clear domain — most commonly, an inside/outside duo or a technology/business duo 
— may find it easier to agree on goals and to be aligned with each other than if there were three 
cofounders who had to be aligned. 

At the same time, the experience of the cofounders at Lynx opens a window into the evolution of a 
team that went fi^om an egalitarian trio to a contentious duo when one of their team moved on. 
Although the team of James Milmo, Javier Pascal, and Doug Curtis had each taken formal titles 
(Javier as CTO, James as president/chairman, and Doug as CEO), they initially formed what James 
called "a triumvirate making decisions. We were a threesome: We could see when two of us agreed 
and the other would relinquish the point, which helped mitigate many problems." 

When Doug left the team, they hired Clark Evans as an interim CEO. In addition to adding 
executive experience to the team, Clark served as the tiebreaker and buffer between James and 
Javier. As James recounted, "Javier and I would have intense debates with each other all the time, 
and the only thing that kept us from going crazy was having Clark around to break it up." After Clark 
phased out of his position, James and Javier could not find a suitable replacement and were left as 
dual (and dueling) decision makers, a configuration that was fraught with conflict, tension, and 
frustration. When they were in the midst of that stage, James explained, "[N]ow, we're extremely 
tired of living by consensus. Without Clark around to mediate, we work through differences endlessly 
and seek a consensus on everything, no matter how long we have to debate an issue. Forcing 
consensus is exhausting! Not only do we have to manage the business, but we also have to manage 
each other. Clark had been the third leg of the stool, but without that third leg now, it has become 
really uncomfortable. We have no structural way of resolving our problems and our decision making 
is breaking down." 

James and Javier fought over nearly everything: how to manage the company, whether to grow 
aggressively or conservatively, and whether they should allocate money toward market testing. The 
constant discord finally forced James to give Javier control over the day-to-day management. As 
James recalled, "Eventually, we had to divide the responsibilities super-clearly just so we could 
fiinction. We decided that it was better to make wrong decisions fast than no decision, especially 
when coming to a consensus was getting harder and harder. I finally relinquished control and said to 
Javier, 'You go where you want, and I'll stop reaching over and taking the wheel or pushing the gas 
pedal when you want the brake. I'll sit on my hands when we disagree.' . . . Frankly, it was a relief" 

Echoing this metaphor — and highlighting the trouble that duos have resolving a split (1-1) vote — 
one founder said, "One thing I am clear about now: Two people at the wheel is the worst way to 
drive. You end up going straight when either a right or a left would be better." Another founder said 
that tensions between him and his cofounder harmed both the startup's exit value and their personal 
relationship. "We finally managed to get an exit for the conpany, but it cost us a lot in our personal 
relationship and, worse, the outcome of the conpany to which we devoted years of our lives. 
Arguably the exit was substantially less than it might have been had we been of one mind or had we 
'fallen into line' with one of us being dictator sooner." 

ONE FOUNDER ON THE BOARD VERSUS MULTIPLE FOUNDERS 

In addition to serving as members of the executive team, many founders also serve as members of the 
startup's board of directors. It is tempting for a founder-CEO to fight to have a cofounder, presumably 



more supportive than outside directors, on the board. Indeed, across all of the startups in my 2006- 
2009 dataset, 67% had at least one founder on the board of directors. Among those startups that had 
not yet raised a round of outside financing — that is, startups in which the decision of whom to put on 
the board still belonged to the founders — 34% of the boards included two or more founders and 40% 
included one founder. But among startups that had raised their third round of financing (the "C- 
round"), things had changed: Only 18%) of the boards included more than one founder {42% included 
a single founder). But on what grounds were the decisions made and with what consequences? 

Most startups begin with an informal "founders' board" that includes all of the founders and takes 
part in major decisions. However, once the founders begin to involve outside parties, such as 
investors, an official board is formed and board seats become more of a zero-sum game, wherein one 
founder's remaining on the board means that another has to leave it. (See Chapter 9, "Investor 
Dilemmas," for more details on the formalization of the board and other effects of investor 
involvement.) During each round of professional financing, a fixed number of seats are allocated to 
founders, non- founding executives, investors, and other outsiders. The number of founder board seats 
is often smaller than the number of founders and shrinks with each progressive round, forcing the 
founders to negotiate among themselves who will serve on the board and very likely creating or 
increasing tensions between them Sometimes — for example, when one founder is the CEO — it is 
clear which founder should remain on the board. But if, for example, multiple (but not all) founders 
can serve on the board, or if a nonfounder is CEO and the founders hold other C-level positions, the 
negotiations about board membership can be contentious. For a founder, becoming a director has 
strong attractions but also high costs. Many founders seek the additional decision-making control that 
comes fi"om serving on the board and feel a deeper affinity to the startup if they can remain a director. 
But serving on the board requires a founder to put in more hours and to play a more complex fiduciary 
role than nondirector founders do. 

After FeedBurner's first round of financing, there were two board seats available for the four 
founders. The team agreed that Dick Costolo, the founder-CEO, should receive one of those seats, but 
had trouble assigning the other. In the end, the team decided that Steve should sit on the board along 
with Dick, mainly because of Steve's involvement in business development and legal issues. At Lynx 
Solutions, all three founders managed to stay on the board through the first two rounds of financing. 
But during the third round, the VCs, wanting to avoid a board that was too big to be effective, 
negotiated with the founding team to reduce the number of founder-directors to two, forcing the 
founders into a difficult negotiation. As founder James Milmo explained, "We were very concerned 
about maintaining the balance of power within the group, not destabilizing things." Rather than leave 
one founder off the board entirely, all three founders continued to attend the board meetings. Although 
they had only two voting seats, they tried to circumvent this constraint by making joint decisions and 
voting as a block. Although this reduced the initial tension over who would be represented on the 
board, it did introduce another recurring source of tension whenever the team disagreed about how to 
vote on the board. 

Even when there seems to be clarity within the executive hierarchy, this clarity can be 
conprorrdsed when non-CEO founders serve on the board of directors. If one founder is CEO but 
other founders are on the board, it is less clear that the CEO has the final say within the executive 
team. Similarly, if the CEO has a subordinate who is also on the board to which the CEO reports, 
their power relationship can be unclear. Having a non-CEO founder on the board can even interfere 
with the board's effectiveness. For instance, if a founder-CEO who is facing challenging cofounder 
issues wants to get the board's advice, a sensitive situation can become even more complicated if the 



cofounder also serves on the board. And when a board wants to give negative feedback to the 
founder-CEO, can it do so effectively — and without causing problems within the executive team — if 
the CEO's cofounder- subordinate is present? 

In the case of Ockham Technologies, even though Mike reported to Jim, both founders served on 
the board, which led Mike to assume that he was an equal when it came to high-level decision 
making. As Jim later recounted, "In hindsight, I would never again put another executive from the 
conpany, other than the CEO, on the board, because you're inevitably going to get into very sensitive 
discussions about how you are going to grow and how you are going to split responsibilities. I would 
have liked to have had those conversations without Mike being a part of that [decision-making 
process]" — even at the expense of losing his most natural ally on the board, his cofounder, and at the 
risk of inperiling their mutual trust. Other founder-CEOs have struck a different balance by having a 
cofounder serve on the board in a limited role. For instance, the board may conduct separate sessions 
that are not attended by non-CEO executives of the corrpany. Alternatively, non-CEO cofounders may 
be given one-year terms on the board or may serve as "observers" who can attend meetings but do not 
vote. 

ROLES AND WEALTH-VERSUS-CONTROL MOTIVATIONS 

We have seen that founders can base their decisions about title assignment, division of labor, and 
decision-making approach on such relatively objective criteria as who has what skills, contacts, or 
passion, who are the idea people, and how all these forms of human capital overlap. However, a 
major factor in role assignment — harder to assess, but often more powerfiil than the more objective 
criteria — is each founder's motivation for launching the startup. Core founders should examine the 
motivations of their potential cofounders to see if they are compatible with their own motivations. 
Looking at two broad sets of motivations — control and wealth — ^we can see configurations that are 
conpatible and others that are likely to be disruptive. 

In a two-founder team, for example, both founders can be control motivated, both can be wealth 
motivated, or there can be one of each. Conceptually, a team of two control-motivated founders 
should be less stable than the others; both will want to be CEO, both will want to serve on the board, 
and both will want final say over important decisions. Much of what undid Blogger in its first 
incarnation was the fact that both cofounders, Evan Williams and Meg Hourihan, wanted control but 
each wanted to go in a very different direction. Even an early hierarchical arrangement can be 
unstable if the cofounders share a control motivation. In the early days of Odeo, Evan agreed to give 
the CEO title to cofounder Noah Glass, but later asserted his control and replaced Noah as CEO in 
order to be able to pilot the startup the way he wanted. 

When both founders are wealth-motivated, they are more likely to be aligned. Their decisions 
regarding roles and decision making should favor whichever arrangements will build the greatest 
value for the startup, regardless of the control inplications. At FeedBumer, the cofounders shared the 
goal of a lucrative sale, which helped align their interests at each stage of the startup's growth and 
increased the chances of their getting to the Rich outcome they desired. 

A wealth-motivated founder and a control-motivated founder can be very well aligned as long as 
the wealth-motivated founder has confidence that the control-motivated founder can build the 
startup's value. 

Potential cofounders need to assess this type of conpatibility before deciding whether or not to 



begin a startup together. Motivational conpatibility does not guarantee success, but inconpatibility is 
asking for trouble. Founders who launch their startup and only later realize that their motivations are 
misaligned will encounter higher tensions within the team at almost every stage of decision making. 
These tensions will be particularly severe regarding the role dilemmas described in this chapter, for 
control of the startup is intimately tied to how titles are assigned and how decisions are made within 
the founding team However, control is also strongly affected by how much equity is owned by each 
founder, an issue examined in our next chapter. As we will see, high-tension negotiations over the 
splitting of roles are often matched by high-tension negotiations over the splitting of financial 
rewards. 

CLOSING REMARKS 

In assigning roles to cofounders, just as in choosing those cofounders in the first place, there is a set 
of natural inclinations — "easy paths" — that founders need to guard against: 

Avoiding Conflict 

When more than one founder wants to be CEO, teams often try to avoid a confrontation by crowning 
multiple decision makers. This is a recipe for delay, lack of accountability, and ineffective decision 
making, especially as the startup grows. While ducking the original conflict may seem to have saved 
the startup from an early inplosion, it actually causes longer-term conflict, endangering the startup's 
fiiture. 

Underestimating "Title Inertia " 

It is surprisingly easy to dismiss as inconsequential the decision about who will be CEO or to 
acquiesce without a fuss when someone (the idea person, or even the founder with the loudest voice) 
crowns himself or herself CEO. In the early stages, everyone is working so hard to make the startup 
succeed that it hardly seems to matter who has which title. However, the CEO has considerable 
symbolic and substantive power and, presumably, the startup won't always be so small and cozily 
egalitarian. The founding team as a group should determine who would be the best initial CEO and 
understand that inertial tendencies will make their decision hard to reverse without major disruption 
if that person cannot keep up with the changing demands of the position or if another founder later 
proves to be the better choice. Most CEOs are reluctant to relinquish their status and power. 

These challenges are particularly acute when, as is often the case, the idea person wants to be 
CEO. Although idea people often seem a natural choice because they possess the passion and vision 
necessary for an early-stage CEO, founding teams should carefiilly assess whether the idea person is 
the best candidate to be CEO on all counts — not just passion and vision — or whether he or she might 
better fit another role requiring those two traits, such as chairman, chief technology officer, or chief 
scientific officer. 

Inflating Titles 

The dangers of inertia likewise apply to non-CEO founders who take C-level titles. Once the startup 



has grown much bigger, many of those founders will not be the best people to fill those senior 
positions. Yet, demoting or replacing "overtitled" founders can be quite disruptive. Given these long- 
term dangers, teams should be more hesitant to assign C-level titles to themselves without 
consciously weighing the pros and cons. 

Wanting Allies on the Board 

When boards of directors are being formed or modified, it is tempting for founder-CEOs to want to 
have allies on the board. Their cofounders are often the first choice. But having more than one 
founder on the board can have long-term costs that outweigh the immediate benefits. For example, 
having a non-CEO founder on the board tends to cause role confusion within the executive team, to 
hinder some board discussions, and to cause even more challenges for the founder-CEO. 

Ignoring Incompatible Motivations 

Founder motivations can have a powerfiil effect on role tensions within the team. If two founders are 
strongly control-motivated, for example, both are likely to want to be CEO, as was the case with 
Evan and Meg at Blogger. Solving this by appointing co-CEOs or creating an ambiguous power 
structure sets up the problems described above in "Avoiding Conflict." In contrast, teams should be 
more stable either when both founders are wealth-motivated (and thus more aligned in their decision 
making) or when a wealth- motivated founder has joined forces with a control-motivated founder 
whose skills and capabilities are up to the challenge of being CEO. Before founding together, 
potential cofounders should assess each other's motivations to understand the potential sources of 
role conflict. 

In Chapter 7, we examine other founding- team problems that are caused by the combination of 
these natural but often misguided inclinations in making decisions about relationships, roles, and 
rewards. 



* Even iconic founding teams who are commonly believed to have been co-CEOs often adopted hierarchical titles within the early 
team. For instance, when Hewlett-Packard incorporated, David Packard was president while Bill Hewlett was vice president. During the 
early days of Google, Larry Page was CEO while Sergey Brin was CTO. Even at ice cream company Ben and Jerry's, where both 
founders "had their name on the door" and on every pint of ice cream, Jerry Greenfield was president. 

* Perhaps the most famous co-CEOs in the technology industry today are Michael Lazaridis and James Balsillie at Research in Motion 
Ltd, which makes the BlackBerry. Commonly believed to have cofounded RIM, Balsillie was actually hired by founder Lazaridis in 1992, 
eight years after the founding. 

* Interestingly, the "lean startup" philosophy for building startups abhors bureaucracy and shuns hierarchy, yet places a premium on 
being able to "pivot" quickly; an experienced high-tech serial entrepreneur points out that "pivoting is very hard to do without a strong 
CEO in charge." 



CHAPTER SIX 

REWARD DILEMMAS: EQUITY SPLITS AND 
CASH COMPENSATION 



Splitting the ownership of a startup is often even more contentious — and may have even more 
dramatic implications — than splitting the roles and titles. For many founders, the main financial 
motivation is the large potential equity upside rather than the paycheck, which is often smaller in their 

cash-poor startups than what they could earn elsewhere.^ Unfortunately, many of our natural 
inclinations about equity splits are wrong or counterproductive, destined to cause problems in the 
long run even when they seem eminently fair, wise, and peacefiil in the short run. Therefore, the bulk 
of this chapter focuses on equity splits, examining the various types of splits, the criteria used when 
splitting, and the central issue of static versus dynamic splits. At the end of the chapter, we will also 
delve into founders' cash compensation, looking at the team-level issue of salary equality and the 
insights we gain into founders' attachment to their startups by examining the "founder discount." 

"WAR" OVER EQUITY SPLITS 

In late 1999, as he was getting Blogger off the ground, Evan Williams negotiated a 60/40 equity split 
with his partner Meg Hourihan. He recalled, "Her expectation was that the equity would be 50/50, 
but I proposed an equity split of 60/40 in my favor; that was the key thing to making it work. I had 
brought the idea and had the more relevant experience and felt that that justified my having a bigger 
percentage. I wanted to get that up-front. I wasn't confident enough to make an argument for 70/30, but 
really wanted 60/40." Meg and Evan settled on 60/40 with a handshake agreement; it wasn't until the 
spring of 2000, when raising a small angel-investor round, that they made the agreement formal and 
added four-year retroactive vesting terms.* Evan said, "The fiind-raising also made everything more 
ofiicial ... the founders' stock agreements became official for the first time; it was the first time we 
paid for lawyers." In the meantime, Evan cared relatively little about receiving a paycheck: "I had 
about $ 10,000 in the bank when we started. I put some of it in the company. I wasn't taking a salary." 
Eventually, Evan would invest all of his own money, max out his credit cards, and work for months 
without pay in order to keep Blogger alive. 

In his next startup, Odeo, Evan agreed to give Noah Glass 70% because Noah had been more 
responsible for the idea and would be serving as fiill-time CEO while Evan would be working only 
part-time. Evan recalled, "I sort of thought of myself as a 'half co-founder." Having had more 
experience with startups at this point, Evan involved "good lawyers" and formalized the equity split 
earlier than he had at Blogger. 

From afar, splitting equity can seem like it should be a very rational process in which each 
founder's value to the startup is matched to his or her equity stake. Evan's approaches to the equity 
splits for Blogger and Odeo seem logical and straightforward, yet both splits burdened the startups 



with agreements that couldn't stand the test of time. Both Meg and Noah left within two years of those 
splits. The break with Meg was particularly painfiil and hard to resolve. She left Blogger at the end of 
2000, after about half of her shares should have vested, and three months later Evan moved to reclaim 
her unvested shares. After six months of legal wrangling, Evan finally settled with Meg. "I just 
decided I needed to be practical and get it over with because it was killing us," he recalled. It was "a 
huge distraction at the worst possible time." He marveled, "I spent more on lawyers in 2001 than I 
paid myself!" 

The problems are visible not only in hindsight. The equity- split negotiation can be among the most 
emotional and visceral events in the development of a founding team, with feelings overwhelming 
rational considerations. Founders often describe their equity-split negotiations as a "war," 
"exasperating," or "stressftil," even when they don't end up experiencing Evan's legal fisticuffs. The 
founders' abilities to fight such battles consttuctively often foreshadow how well they will be able to 
handle other upcoming sensitive issues. 

WHEN TO SPLIT 

Founders can choose to split the equity either at the time of founding or later.* In my dataset, 73% of 
teams split the equity within a month of founding, a striking number given the big uncertainties early in 
the life of any startup. The founders of UpDown, for example, split early in the life of the startup, 
before they knew each other well and could judge each other's abilities and commitment. In fact, their 
initial split severely underestimated one founder's contributions and overestimated another's. Such 
mistakes can be very painful to correct. Researchers have shown that initial psychological "anchors" 
have a powerful effect on subsequent negotiations and final outcomes, making it hard for even experts 
to overcome the inertial effects of anchors.^ The anchoring effect (and legally binding nature) of 
UpDown's early one-page equity-split agreement caused major tension within the team when one of 
the founders wanted to renegotiate the split a few months later. 

In contrast, putting off the split for several months or more can give the cofounders a chance to 
learn whose skills and connections will contribute the most to the startup, how those contributions 
change as the startup's strategy and business model change, how well each founder gets along with 
the others, how committed each founder is to the startup, and more. Founder-CEO Vivek Khuller, for 
exanple, pushed his Smartix team to wait as long as possible — until they were forced to split while 
raising their first round of financing. "You shouldn't exercise an option until you need to," he 
explained. "There's always more you can learn about the venture and your team and that might change 
how much equity each person should get." Holding off on the split may also provide a strong 
incentive for the founders of a young startup to conttibute and to prove themselves, rather than free 
riding* once the split has been determined.-^ 

In founding a startup, almost every benefit comes with a risk. Holding off on the equity split in 
order to be more informed and to keep the cofounders motivated can also cost the team an opportunity 
to attract another cofounder. The Smartix team, for example, was seriously disadvantaged by the 
cofounders' ignorance about how sports venues worked and a lack of industty contacts. Early in the 
evolution of the team, though, they had within their grasp a potential cofounder, David, who had deep 
industry experience and whose contacts had yielded some inportant early meetings for them. But 
David ended up pursuing a more atttactive job opportunity. Vivek later concluded that by deferring 

the equity-split negotiation, he lost the chance to offer David a large stake to join the team"^ 



Furthermore, equity- split negotiations are usually much calmer when done before the stakes 
become very high, which generally means before any financing has been secured and an objective 
valuation has thus been placed on the startup. Negotiating when millions of dollars of financing are on 
the line leads to very different dynamics and can make it very hard for the cofounders to agree.* 



Figure 6. 1. Reasons to Split Earlier versus Later 

Figure 6. 1 summarizes the factors weighing in favor of splitting early versus late. 

The cofounders themselves may differ about when to split. Founders whose major contributions 
come early in the life of the startup, such as the founder who had the idea or the founder who 
contributed the most seed capital, may want to split earlier than those whose major contributions are 
yet to come, such as the technical lead who will develop the product once the other founders have 
developed the product specifications. Ideally, the equity split will approximate each founder's long- 
term level of contribution, but judgments about those levels of contribution will naturally be affected 
by the founders' early experience with each other. During the ups and downs of the startup's 
evolution, it is inevitable that the importance of each founder's contributions will wax and wane, and 
so may their idea of the best time for the equity split. (When founders disagree about whether to split 
early or late, they might be able to resolve the disagreement by using the dynamic approach discussed 
later in this chapter.) 

Sometimes a founder can use these fluctuations in the team members' relative importance to 
improve his or her negotiating position. For example, when Frank Addante and John Bohan started 
L90 — a merger between Frank's company, ReaXions, and John's much more valuable company, 
Adnet — John offered to give Frank half a million shares, out of 15.5 million shares outstanding, 
reflecting the fact that Frank was also far less experienced than John was. But Frank was hesitant to 
make a firm decision about ownership at that early point: "What I said was, 'Why don't we see how 
this thing goes, and we can have the discussion six months fi^om now.' So we didn't decide what the 
equity breakout would be up-front. I decided to wait until later to see what kind of value I [would] 
add and what kind of value [John and the other enployees would] add." Six months later, an angel 
investor offered to invest $2 million for 20% of the company, for the flrst time placing a value on the 
startup ($10 million), and Frank took that opportunity to carve out his share. He asked for and 
received 1 million shares out of a total of 20 million shares outstanding. In this case, waiting earned 
Frank paper gains of about $250,000.* Likewise, a junior cofounder who is confident in his or her 
own skills may beneflt disproportionately from delaying the split. 



Split earlier 4 



> Split later 



• Attract key players who need 
equity incentive 

• If already worked extensively with 
cofounders in another startup 

• Negotiate calmly before you're 
under pressure to split 



• Learn about cofounders' contributions 

• Solidify startup's strategy and business model 

• Solidify roles 

• Learn about cofounders' commitment; 
strengthen incentives 

• Avoid continual renegotiations as things 
change 



CRITERIA FOR EQUITY SPLITS 



If cofounders have decided to negotiate the split, what are the criteria they should (and often do) 
consider? On the one hand, there are no "right" answers and no objective criteria that can be used; the 
outcome is fully subject to negotiation between the founders. On the other hand, from my quantitative 
analyses and in-depth examination of actual equity splits, at least four sets of criteria emerge that can 
help sharpen the negotiation and increase the chances of crafting a sustainable agreement. Those 
criteria are past contributions to the startup, opportunity cost, fiiture contributions to the startup, and 
founder motivations. 

Past Contributions 

First, each founder's equity stake is often based, at least in part, on how much — relative to the other 
founders — ^he or she has contributed to building the value of the startup so far. (Outside of questions 
of fairness, the team's ongoing relationship often benefits from rewarding such "bygones.") The extent 
of a founder's contribution can depend on when the team decides to do the splitting — at the time of 
founding or several months later — ^but even very early on, many founders have contributed at least an 
idea or intellectual property on which the startup is based and/or seed money to fimd the startup. 

The Idea Premium 

All other things being equal, do idea people deserve an idea premium — -a greater equity stake than 
other founders? Many academics and non-idea founders argue that "ideas are cheap; execution is 
dear." However, we saw that Evan Williams certainly placed a value on either having the idea, as he 
did for Blogger, or helping to shape it, as he did for Odeo. In my field research on founders, I have 
come across many who share Evan's attitude toward the importance — even the sanctity — of the idea. 
Frank Addante demanded more equity in Zondigo, his fourth startup and his first as founder-CEO, in 
part because he had provided the idea. Frank explained how his three-founder team had agreed to 
give him most of the equity: "I had realized in past ventures that the value I was contributing wasn't 
matching my equity amount. I had gotten advice from other folks that there should be a premium for 
the idea. . . . My cofounders agreed. Based on my research and advice from others, I offered them 5% 
each. . . . They were fine with that." Phuc Truong described how the team for his first startup, 
Crimson Solutions, thought about equity: "We had a short discussion about splitting the equity and we 
decided to split it 50-25-25. Seth and I both thought that Wellie deserved more because he had the 
idea and had organized everything. . . . Without him there would be nothing. So it felt right that he'd 
get much more than us." In UpDown, Phuc's next startup, the team negotiated an early agreement in 
which they awarded Michael, the idea person, a 5% premium 

Other founders, like James Milmo, believe that their ideas should be rewarded not with more 
equity, but with more immediate equity. James e?q)lained that he agreed to split the equity in Lynx 
equally with his cofounder, but, "I told him, 'I want you to [have to] vest, but I don't want to [have to] 
vest. It was my idea and I invited you to join the conpany. I already earned my stake — my invention 
and patent are worth a lot — but you have to earn yours by working for two years." 

Not all ideas are equally mature or valuable. Some ideas identify a market need but are far from 
proposing a solid solution; cofounders are needed to refine and develop it. Other ideas are backed by 
patents, years of research, and a solid business model. One can also draw a distinction between an 
idea and an idea for making money from that idea. For exanple, Steve Wozniak's father felt that Steve 



Jobs shouldn't receive equity because "he hadn't done anything,"^ whereas Wozniak had actually 
invented the personal computer on which the startup was built. On the other hand, Jobs sometimes 
indicated that he believed he deserved more equity than Wozniak since without Jobs's insistence on 
starting the company, Wozniak would probably still be an engineer toiling at Hewlett-Packard. 

Across this spectrum of "ideas," is there an idea premium, and, if so, how large is it? Is Crimson 
Solutions 's 25% typical, or is UpDown's 5% more the norm? To examine this quantitatively, I 

included a survey question that would identify which cofounders were the idea people,^ along with 
other detailed questions about the founding team's characteristics, and created regression models to 
analyze whether idea people received an idea premium.^ My analyses showed that there is indeed a 
statistically significant idea premium: All other things being equal, idea people receive 10 to 15 more 
percentage points of equity than do non-idea people.^ In addition to rewarding this past contribution, 
the idea premium may also be a recognition that the idea person might be more likely to contribute 
important additional ideas in the future and to make other contributions that are linked to having been 
the idea person. 

For teams where it is clear which founder had the idea and that the idea is valuable, the team may 
be risking resentment if it does not award at least a small idea premium to its idea person. Even if the 
idea person is initially willing to forgo the premium in order to avoid putting sand under the saddles 
of the non-idea cofounders, it may be at the cost of putting sand under the idea person's saddle — i.e., 
introducing friction that grows over time. This sand will continue to chafe as the idea person's 
resentment grows over not being rewarded for having had the idea, thus creating another source of 
tension within the team 

Capital Contribution 

Figure 6.2 shows the distribution of capital contributed by each founder in my dataset. Of the 
founders, 41% did not contribute any capital. Of the other 59%), the majority contributed between $1 
and $25,000, but 11% contributed more than $100,000. At the team level, in 42% of ventures, all 
founders on the team contributed some amount of capital. 

In 38%) of teams, the amounts of founder capital differed within the team. My regression analyses 
showed that the more capital a founder had contributed, the larger that founder's equity stake. ^ 
Ockham's founding team actually used the differing amounts of seed capital as the sole determinant of 
equity stakes. Founder-CEO Jim Triandiflou recalled that, at first, the idea was that each founder 
would contribute one-third of $150,000, but neither of his cofounders wanted to give that much, so "I 
said . . . put in whatever you want, I'll make up the difference. ... In the end, I had about 50%o, Mike 
had about 30%), and Ken had about 20%o. That's how [the equity split] ended up — simply because of 
the dollars." 

At one level, founder capital is a very tangible contribution to the startup, allowing the team to rent 
an office, pay the phone bill, and make presentations around the country. Yet, differences in the 
financial capital contributed by each founder may sinply reflect the differing ability of each founder 
to contribute money. However, at a deeper level, a founder's capital contribution may also indicate 
his or her commitment to and confidence in the startup. It may thus foreshadow which founders will 
make the biggest contributions to the startup, all other things being equal. In Ockham's case, although 
Ken had generated the idea and had worked with Jim even before they added Mike as a cofounder, 
when it came time to actually commit money to the startup. Ken was the least enthusiastic and 



contributed the least capital. This decision anticipated his ultimate departure from the startup; when 
Jim and Mike quit their jobs to commit full-time to Ockham, Ken admitted he enjoyed his consulting 
job and, with the birth of his first child, was unwilling to make the leap. 




Opportunity Cost 

Some founders are involuntarily unemployed when they begin discussions about joining a founding 
team. Others, such as Ken Burows at Ockham, are frilly employed in high-level positions which they 
enjoy and which give them a financial security they find hard to forgo. For the first group, the 
opportunity cost of joining the startup is low, but for the others, the opportunity cost is high and the 
rest of the team may have to make the startup opportunity more atfractive to entice the potential 
cofounder. Sometimes, the attraction is the chance to be a founder, to get a more senior titie, or to 
play a more cenfral role. Other times, it may be the chance to earn a large equity stake. 

In forming the Smartix team, Vivek Khuller identified Saurabh Mittal, one of his classmates, as 
someone who could contribute greatly to the startup. Vivek explained, "During our discussions, I 
learned that Saurabh was already working on two other high-potential startup ideas. But he expressed 
interest in the Smartix concept so I decided to get him involved." Vivek and his team conducted a 
lengthy equity-split negotiation that culminated in Saurabh receiving the most equity (27%) after 
Vivek (35%), in large part to attract him away from his other potentially lucrative options and to 
conpensate him for working without a salary. 

Similarly, when Phuc Truong joined UpDown, he had already worked on a successfiil startup and 
had several lucrative software development jobs. When the team discussed equity splits, Phuc's point 
of view was that, whereas the other cofounders "would only be giving up a social life at business 
school," he would sustain a tangible — even quantifiable — loss in income by joining UpDown fiill- 
time. He would therefore need either a substantial salary or more of the equity as compensation. 



Future Contributions 

A founder's future contributions, although often the hardest factors to evaluate, are in many ways the 
most important for determining equity stakes. "Remember," warns one entrepreneur, "when the pie is 
split, 95% of the work required for success remains in the ftiture." Often, a founder's potential to 
contribute to the startup can be estimated by considering his or her background in combination with 
his or her commitment level. 

Researchers have already observed that specific founding experience is more valuable for startup 
growth than are overall work experience and educational human capital.^ Serial entrepreneurs are 
commonly expected to have stronger human and social capital (and possibly even greater financial 
capital) and are therefore expected to contribute more to building the startup's value. My quantitative 
analyses with Thomas Hellmann found that a prior founding experience was associated with a serial 
premium of 7 to 9 percentage points more of an equity stake. Yet, after controlling for prior founding 
experience, status as an idea person, and capital investment, a founder's greater overall prior work 
experience — any work in any company — did not lead to a larger equity stake. Thus, specific founding 
experience is indeed much more important for equity splits than is general human capital. (As 
discussed in Chapter 7, different founding roles and titles are also associated with greater equity 
stakes.) 

The level of fiiture contribution is also presumed to depend on how much time a founder can or 
will commit to the startup. Full-time founders tend to receive larger equity stakes than part-time 
founders. For instance, in her third startup, serial entrepreneur Tracy Burman wanted to cut back on 
her workload to spend more time with her young children. She therefore negotiated two terms with 
her cofounders. First, rather than be the CEO, as she had been in the past, she would be the COO, 
reporting to the CEO. Second, she would work 80% of the time that her cofounders would work, 
receive 80%) of the equity stake that she would normally have received, and earn 80%) of the other 
founders' salaries. 

Founder Motivations and Preferences 

As with most of the dilemmas examined in this book, the founders' motivations play a central role in 
the equity split. Founders with strong wealth motivations will place a high premium on maximizing 
their financial gains fi^om the startup. As described in Chapter 7, other motivations will lead other 
founders to prioritize a particular role, title, or lifestyle flexibility. 

Related factors that may influence a founding team's equity split are as follows: 

• A founder's degree of risk aversion, confidence in his or her own abilities, and confidence in 
the startup's prospects will affect how much priority he or she places on gaining every 
additional percentage of equity versus gaining additional cash compensation fi"om the startup.* 

• Tolerance for conflict will affect a founder's willingness to engage in tension-filled 
negotiations over each type of financial reward. 

• Prior relationships among the founders may influence the split, as described in Chapter 7, 
"The Three Rs System." 

Figure 6.3 summarizes the core factors that should affect equity splits as well as some of my research 
findings about how much of a difference each factor actually makes. All of these factors deserve 



further study by researchers and careful attention from founding teams. 



EQUAL VERSUS UNEQUAL SPLITS 

It seems unlikely that any two founders could be making exactly equal contributions to the value of the 
startup, have the same opportunity costs, and have equal motivations. Thus, it would follow logically 
that if equity splits are meant to reflect the founders' relative contributions and value to the startup, 
equal splits should be rare or nonexistent. Such is not at all the case. As shown in Figure 6.4, 33% of 
the teams in my dataset split equity equally."'' Why did they make this decision? Was it a good 
decision? Did it make the team more stable? Is it associated with higher or lower valuations when 
raising outside financing? 

1. Past Contributions: How much has the founder contributed to building the value of the 
startup so far? 

a. Idea Premium: Founders who contribute the original idea on which the startup is based 
have made a unique contribution to the venture.* 

b. Capital Contribution: Founders who have made larger contributions to the startup's seed 
capital should see a proportionate increase in their equity ownership. 

2. Opportunity Cost: What are the founders sacrificing in order to pursue the startup? 

3. Future Contributions: Most of the work required for the startup to be successful will come 
in the future, but these contributions can be hard to anticipate. How much can each founder 
be expected to contribute to the value of the startup down the road? 

a. Serial Founders: Members of the founding team who have previously led a startup to a 
successful exit can be expected to contribute more human and social capital down the 

road.^ 

b. Level of Commitment: Founders who are committed full-time to the venture can be 
expected to contribute more value. 

c. Titles: The official positions of the members of the founding team have been shown to 
influence equity splits, with CEOs receiving a substantial equity premium. ^ 

4. Founder Motivations and Preferences 

a. Wealth Motivations should lead founders to prioritize larger equity stakes. 

b. Risk Aversion and Optimism will affect how much priority a founder places on gaining 
equity versus cash compensation. 

c. Tolerance for Conflict will affect a founder's willingness to engage in negotiations. 

d. Prior Relationships can affect expectations about equity splits (see Chapter 7). 

Figure 6.3. Factors for Deterinmmg Equity Splits 



Huge gap 




Figure 6.4. Difference in Equity Stakes between Founder with Largest Stake and Founder with Smallest Stake 

A Threshold of Psychological Pain versus Financial Gain 

Thomas Hellmann and I analyzed my data to assess possible explanations for the prevalence of equal 
equity splits. Assuming that negotiating an equity split is inherently tense, so that most people would 
have a strong motivation to avoid it if that could be justified, we wondered whether there was a 
threshold of intrateam difference that separated equal-split from unequal-split teams. That is, if 
cofounders differ enough in the value of what they bring to the startup — be it experience, ideas, 
investment capital, or fire in the belly — the team will be willing to engage in a tension-filled 
negotiation over equity, because a simple equal split would seem too inequitable. Below a threshold, 
however, the gains from engaging in such a negotiation would not be worth the problems introduced, 
and the team would be inclined to split the equity equally. Tim Westergren of Pandora Radio, for 
example, was warned that negotiating over equity, rather than simply splitting equally, would 
"introduce sand under the saddle," causing friction within the team After pondering that advice, Tim 
decided that whatever extra equity stake he might be able to insist on wouldn't be worth the potential 
problems, and he agreed to split equally. 

Hellmann and I gauged how similar or dissimilar a team's founders were in value by comparing 
two aspects of the founders' backgrounds — prior founding experience (i.e., he or she was a serial 
entrepreneur) and prior years of work experience — and two of their early contributions to the startup 

— ideas or intellectual property and seed capital. Our quantitative results indeed showed that "close 
to equal" founding teams tended to split equally, while more heterogeneous teams tended to split 



unequally. This is consistent with the "sand under the saddle" wisdom that causing friction within 
close-to-equal teams can be counterproductive, leading such teams to split equally and avoid the 
high-tension negotiation. Our finding is also consistent with the fact that roughly equal contributions 
are cognitively difficult for the founders to distinguish, so that close-to-equal teams will find that an 
equal equity split feels more sensible. 

"We're a Team" 

Some founders argue that an equal split sends a good signal to everyone on the team. One founder 
explained that it showed that "we were more concerned with the success of the enterprise than our 
own personal wealth accumulation; that we had the presence of mind to realize the success of the first 
would ensure the latter. . . . We're all starting from the beginning, we're all taking the same risk, 
we're a team. If we're successfiil we'll all 'get enough' without trying to jockey for equity position 
this early in the game. Jockeying for position this early in the game does nothing for the effectiveness 
of teamwork." Of course, showing unity and avoiding conflict can be overlapping motives. 

Financing Outcomes: Quick Handshakes versus Thoughtful Equal Splits 

Some teams arrive at an equal split after carefiilly concluding that the founders have similar 
backgrounds, resources, or levels of commitment and can therefore be expected to contribute similar 
amounts to the startup. It is a very different matter, though, when the team prefers to avoid negotiating 
the split and defaults to a quick handshake and an equal split. Indeed, we found that equal splitters 
tended to spend much less time negotiating the split than did unequal splitters. Of the equal splitters, 
60% of the teams spent a day or less negotiating the equity split, while only 39% of the unequal 
splitters made their decision that quickly. 

We categorized the equal-split teams who had spent a day or less negotiating their split as "quick- 
equal" teams and those who had spent more time on the negotiation as "slow-equal" teams. 
Interviews with founders suggested that there were qualitative differences between quick-equal and 
slow-equal teams. For instance, one serial entrepreneur observed that "a quick-equal split is a 
symptom of a founding team that doesn't have any experience or methodology for making the founder 
split so they just say screw it and do 1/N."* Another founder echoed that "a quick handshake is a 
symptom of inexperience and/or lack of real preparation, or real glue to the idea, before going on the 
'tough journey' " We'll see examples later in this chapter of founders who regretted making an equal 
split. 

To assess quantitatively whether quick-equal teams and slow-equal teams fared differently (due to 
underlying differences between such teams, rather than due directiy to how they split the equity), 

Thomas Hellmann and I focused on the first round of financing. Our analyses showed that, 
controlling for a wide variety of differences across teams and startups, quick-equal teams received 
significantiy lower valuations than slow-equal teams and unequal-split teams, lending credence to 
anecdotal evidence that there is a real difference between teams who engage in a serious dialogue 
about the split and those who avoid such a discussion and rely on a quick handshake. The handshake, 
for all its cultural resonance as a symbol of trustworthiness and echoes of a golden age "when you 
could do business with a handshake," is often a symptom of a founding team's weakness. Many VCs 
report that they do examine the team's equity split before investing and that a problematic split could 



have important consequences. One explained, "We spend time during our diligence process 
understanding how the split came to be" (e.g., whether it had resulted from a quick handshake); 
whether the split had "led to suboptimal behavior," indicated weaknesses within the team, or 
introduced additional tensions within it; and whether those issues should affect the startup's 
valuation. 

51% versus 50% 

Some founders, particularly in a two-founder team, will argue for a 51% equity stake, and sometimes 
the other partner is quite willing to go along with that. In one of his earliest startups, serial 
entrepreneur Frank Addante agreed to such a split. When he and his cofounder, Gary, started 
ReaXions, Gary fiinded the startup but was otherwise a silent partner, while Frank was to receive a 
salary for his fiill-time work. Gary insisted on receiving 51% of the equity; since he was putting up 
the money, he wanted control over how that money was spent. As Frank explained, "49%) versus 51%o 
meant we had a similar amount of equity, but if a major decision needed to be made, it was his." 
Thus, the additional 1% of equity can have a substantial impact on decision- making control. 

Frank was willing to give Gary 51%), but in many founding teams, that additional 1% can cause 
major problems. It may suggest that one cofounder is overly focused on control instead of on crafting 
an effective partnership, or it may foreshadow a dysftmctional team dynamic. The 51% share can also 
provoke resentment. One founder complained, "My partner, who came up with the idea, controls 51% 
of the company, based on an early conversation we held. However, I am providing the initial ftmding 
and have done most of the work as of late and I feel a 50-50 split is plenty fair." (This example also 
makes a good case for the kind of dynamic splits described later in this chapter.) 

Founders should realize, though, that even the "absolute control" seemingly granted by a 51%) 
equity stake is not absolute and may not last. For one thing, a 51%) equity stake does not guarantee 
direct control of all decisions. The most important decisions, such as sale of the corporation's assets, 
are left to a vote of shareholders; here, the founder with a 51%o share does have absolute control. But 
the next tier of important decisions is in the hands of the board, which is elected by the equity holders 
(who may also have seats themselves). As the number of directors increases, the equity stake 
necessary to elect a director decreases. Most importantly, the board elects the GEO, who controls the 
third tier of decisions, including oversight of day-to-day operations. 

In any case, for teams planning to raise at least one round of outside financing, the power of 51%) 
equity will be only temporary. As soon as the team takes some outside financing, the 51%) stake will 
shrink to less than 50%o and that founder will have to build a coalition with other shareholders to have 
his or her way. * (In addition, as described in Ghapter 9, investors often receive securities that have 
more rights than does the simple "common equity" received by founders.) Nike cofounders Phil 
Knight and Bill Bowerman, for exanple, originally split the equity 51%-49% in Knight's favor and 
Knight resisted giving up equity, even to a crucial early employee, in order to keep control of the 
firm By 1971, however. Knight was desperate for fiinding and accepted $200,000 from a private 
offering. In return, he agreed to give up 35% of the company and with it his controlling interest. 

FORMAL VERSUS INFORMAL SPLITS 

When the founders split the equity, they can either keep the agreement informal or commit it to 



writing. As shown in Figure 6.5, the longer the cofounders negotiate to reach agreement on the split, 
the more likely they are to commit the agreement to writing, either on their own or with a lawyer. 
Teams who split quickly (within one day) are much more likely to stick with a verbal agreement 
(22% of the teams, compared to 11% for the teams who spend longer negotiating) and less likely, if 
they do formalize the agreement at all, to do so right away (41%) versus 54% of longer-negotiating 
teams). 

100% 
90% 
80% 
70% 
60% 
50% 
40% 
30% 
20% 
10% 
0% 





1 day or less 


2 days or more 


□ Formal 


41% 


54% 


□ Informal, later committed to 
writing 


37% 


35% 


■ Informal only; never committed 
to writing 


22% 


11% 



Figure 6.5. Time Spent Negotiating the Split versus Formality of the Agreement 

Two factors might affect this linkage between negotiation length and formal agreement. First, when 
the agreement is straightforward — ^with no complex or subjective contingencies and no intellectual 
property issues — the team is more likely to be able to agree quickly and less likely to feel the need to 
capture the details of the agreement in writing. (On the flip side, teams who are in a rush to get to an 
agreement may craft simpler agreements.) Second, teams who try to avoid the tension of negotiating 
the split (by doing so as quickly as possible) may also be more likely to try to avoid the tension of 
committing the agreement to writing, which can involve reopening some of the issues. 

While the advantages of a formal agreement are obvious, there can also be disadvantages. For 
instance, overly rigid contracts can lock parties into arrangements that don't allow for adjustments as 
circumstances change, and overly detailed contracts can undermine trust by preventing spontaneous 
displays of good intentions. When formalizing the agreement, founders should also anticipate ftiture 
tax or legal issues that could be avoided if handled appropriately.* 

STATIC VERSUS DYNAMIC SPLITS: THE PERILS OF SETTING THINGS IN STONE 

Setting the early equity split in stone, without allowing for changes, is one of the biggest mistakes 
founders can make. With their confidence in their startup and themselves, their passion for their work 
and their mission, and their desire not to harm the fragile dynamic within the nascent founding team. 




cofounders tend to plan for the best that can happen. They assume that their early, high levels of 
commitment will last long into the future, rather than waning as the challenges of founding begin to 
sap their passion for the idea and for each other. They assume that no adverse events will change the 
conposition of the team. They also tend to take a very short-term view of the factors that should affect 
equity splits. They assume that the tasks that they are performing during the early stage of startup 
development are the same tasks that will be performed during the next and very different stages. They 
assume that their skills will remain as valuable to the startup as they are right now. They overestimate 
the amount of value that they will build in the first months compared to the value they hope to build 
over the subsequent years, and thus overweight their past contributions compared to the future 
contributions that will be required of them Each founder places more value on his or her own 
contributions than on the contributions of the other cofounders, knowing the cost and extent of his or 
own efforts in a way that he or she cannot know the cost and extent of others' efforts. 

But such a best-case approach is hazardous. Uncertainties abound. At the company level, founders 
learn about the flaws in their initial plans and adjust the startup's strategy, business plan, and business 
model. Professor Scott Shane reports that "almost half (49.6 percent) of new firm founders indicated 
that their business ideas [had] changed between the time they first identified them and the time when 
they were surveyed about them."^^ Such adjustments can cause major changes in the obstacles that the 
startup faces, the skills needed to address those obstacles, and thus the roles that each founder (or 
perhaps a new founder or a nonfounder) will have to play in building the startup. 

At the individual level, as the strategy and business model shift, the skills of some founders 
become more important than the skills of others and roles often shift. As each founder learns about the 
demands of building a startup, reflects on his or her motivations, and sees how well his or her 
abilities address the startup's needs, his or her commitment to the startup may change. The founders 
also come to understand each other's abilities and commitment at a far deeper level than was 
possible at the beginning. Yet founders tend to overestimate how much value they will build during 
those early days, which can cause even bigger problems when a cofounder's contributions wane later 
on. 

A founder's personal life may also affect his or her commitment and contributions. At Ockham, all 
of the founders were aware of the imminent arrival of Ken's first child. However, even Ken was 
unsure how this would affect his willingness to quit his full-time job and focus on building Ockham 
Extreme and unexpected health problems can catch all parties by surprise. For instance, while 
Microsoft was still a private company, cofounder Paul Allen was diagnosed with Hodgkin's 
lymphoma, which caused him to quit the company, leaving Bill Gates as the sole active founder 
during the crucial three years before it became a public company.* 

In such ways, even the most comfortable equity split can be thrown into disarray. For instance, 
when Robin Chase and her partner, Antje, founded the car- sharing startup Zipcar, they agreed with a 
quick handshake to split the equity 50-50. The team believed it had avoided destructive tension over 
the equity split and could now focus on building the startup. "We shook across the table, 50-50," 
Robin recalls, "and I thought 'great' " Robin had heard about other teams that had faltered because of 
tough equity-split negotiations, and she breathed a sigh of relief that she and Antje had avoided such 
problems. Robin poured her heart and soul into the startup, making major contributions to its growth, 
and was fiilly expecting Antje to do the same. Antje, however, remained at her full-time job and, by 
the summer, was expecting her second child. Robin wondered when her partner would be able to 
become more involved, but, in the end, Antje never joined full-time. Knowing that Antje still owned 



the same percentage as she did ate away at Robin, who later reflected, "That was a really stupid 
handshake, because who knows what skill sets and what milestones and what achievements are going 
to be valuable as you move ahead. That first handshake caused a huge amount of angst over the next 
year and a half."^'^ Eventually Antje left the company altogether while continuing as a shareholder. 

The cost to fix such problems can be very high, ranging from Robin Chase's "angsf to more 
tangible financial costs. At govWorks.com, founders Kaleil and Tom had a cofounder, Chieh, who put 
up $19,000, worked "after hours" for five months (he had kept his day job instead of joining 
govWorks fiill-time), and then dropped out. When the remaining cofounders were about to close their 
first round of financing, their potential funder, Mayfield, was not willing to close until Kaleil and 
Tom bought Chieh out and reclaimed his equity. The VCs were willing to do a $410,000 "sweetheart 
deal" to facilitate the buyout. However, Chieh wanted $800,000. Amid the pressure to close the 
round, Kaleil and Tom ended up settling with Chieh for $700,000, making up the $290,000 out of 
their own pockets. Kaleil felt he was "being extorted." Although the risks of this kind of outcome 
are real, teams often fail to address them proactively. In my dataset, half of the teams had neglected to 
include any dynamic elements (vesting, buyout terms, and the like) in their equity agreements, 
sentencing themselves to the same risks faced by the Zipcar and govWorks.com teams. 

A Teii^late for Founders 

How should founders deal with such developments? In short, by assuming when they do the initial 
split that things will change, even if the specific changes cannot be foreseen, and therefore structuring 
a dynamic equity split rather than the static splits used at Zipcar, govWorks, and many other startups. 
As important as it is to get the initial equity split right — by matching it as closely as possible the 
founders' past contributions, opportunity costs, fiiture contributions, and motivations — it is equally 
important to keep it right; that is, to be able to adjust the split as circumstances change. 

The UpDown team provides a good example of the need for a dynamic agreement and a template 
for discussing forward-looking issues. They first split the equity in November 2006. Up to that point, 
all four founders (one of whom. Warren, had become a dropout cofounder who would no longer be 
contributing to the startup) had worked only sporadically on the business plan and there was no 
fimding in sight. The three remaining cofounders therefore decided to ignore past contributions (with 
the exception of a small idea premium for Michael), assume that they would make equal fiiture 
contributions to the startup, and split the equity equally except for Michael's 5% idea premium 

A month after the team had made this agreement, Michael entered talks with an interested angel 
investor and was motivated to increase his work effort dramatically. At the same time, Phuc was busy 
with his own consulting jobs and Georg was still on an extended vacation with his wife and two 
young children. After a few weeks, Michael worried that his team members' current motivation and 
contribution levels might foreshadow their fixture contributions and thus set out to develop what he 
felt would be a more accurate template to use for a new equity split, one that would award him more 
equity for building more value. A simplified version of the template is shown in Figure 6.6. 

Overall, with the addition of the dynamic elements described below, Michael's template can 
provide a model for other teams grappling with how to split their equity. It breaks down the 
qualitatively different stages of startup development into separate phases, provides a structured way 
to discuss and weight the importance of each phase for building the startup's value, and provides an 
opportunity to plan and evaluate the various tasks that are being or will be performed by each 



founder. Michael's cofounders disagreed vehemently with several details in his tenplate (e.g., the 
overweighting of Phase I, the list of contributions by Phuc, and changes to the Phase III contributions), 
but the tenplate provided a structured approach to discussing a variety of complex factors, thereby 
facilitating the team's renegotiation. 

A structured approach can also serve as a check on the natural tendency to overemphasize tangible 
factors at the expense of intangible factors which may have more impact on team dynamics and the 
success of the startup. For founders — and for researchers! — ^past contributions are usually easier to 
assess than future contributions, cash contributions are easier to assess than contributions to honing 
the idea, and skills are usually easier to assess than commitment and motivation, but teams should 
seek mechanisms that enable them to discuss and balance this wide variety of factors. By developing 
the template together, a founding team may also be able to shift its focus away from wrangling over 
specific numbers and toward a more productive process for agreeing on criteria, arriving at 
weightings, and negotiating tasks and responsibilities. 





Phasf I: Oil ZOOd-jM 2007 


Phjse U: Feb-. 


Mjv 2007 


P/jufa- ///: Jun-JiJ 2007 




Fnutider 


40% 
C-tinlrilHitiitn 


Etfuity 


Contriliition 




Contrilnitiftn Eijuity 


Total Veightcd 
F.tjuity Stjke 


Niidiacl 


Idcntilk.i(ion 
of opponunity; 
define business 
strategy. Kind- 
raising 


60% 


Investor 
management 


33'/j% 


Same 33'/!% 


45% 


Geofg 


Define business, 
iTurkct 
strategy, etc. 


28% 


Marketing, 

product 
management 




Same 33Vj% 


2S% 


Phuc 


Initial mock site; 
dcline business 
strateg}- 


10% 


Site 

improvement 


J3'/j% 


Same ^ >'/>•><■ 


26% 


Warren 
TomI 


Dcline business 
stratcg>'; 
financial plan 


2% 
100% 




100% 


100% 


1% 
100% 



Figure 6.6. UpDown's Template for Splitting Equity 

Recall that the UpDown team negotiated its equity split twice in order to account for the varying 
contributions of each founder over time and also used the template to do some forward-looking 
planning, but even that wasn't enough. After the team had agreed on a new split in February 2007 and 
had worked through the summer, both Georg and Michael applied for work visas, which would 
enable them to quit school and work fiill-time on UpDown. In the 50-50 visa lottery, Michael was 
granted a visa but Georg was not, making it impossible for Georg to be involved in the startup fiill- 
time. They hadn't thought of this during their equity-split renegotiations and now had to work out the 
equity split one more time, again heightening the tension within the team. Anticipating such changes, 
in the ways discussed below, would have helped them avoid fijrther tensions by structuring the initial 
agreement to adjust to those changes. 

Terms, Contingencies, and Trust 

Dynamic equity agreements can make use of a number of approaches, such as buyout terms (in which 
a founder's stake might be bought on prenegotiated terms by the startup or by other founders) and 



vesting schedules. At their core, however, all of these structures are geared to deal with the 
uncertainties inherent in startups. 

Inspired by a famous quote from former U.S. Secretary of Defense Donald Rumsfeld, my colleague 
Deepak Malhotra has developed a general way to categorize the different types of uncertainty 
involved in contracts as "knowns" (whose outcomes are already known or are ensured), "known- 
unknowns" (scenarios for which one can anticipate the occurrence but not the outcome), and 
"unknown-unknowns" (whatever complete surprises the fiiture holds in store). These three types of 
uncertainty can be addressed by terms, contingencies, and trust, respectively. 

Founders' equity splits involve knowns, such as how much capital each founder has contributed 
and who owns the patents, which can be addressed using standard contractual terms. For example, 
Zondigo and Crimson Solutions dealt with the fact the one of the founders had had the original idea by 
giving him a larger share of equity; Lynx handled the same known by means of a vesting arrangement. 
Ockham dealt with its founders' differing cash contributions by splitting the equity in like 
proportions. 

Known-unknowns can be addressed by using contingent provisions that outline how the equity split 
should change for various worst-case, expected-case, and best-case scenarios. The Ockham team, for 
example, did not know what Ken would do once his new child was born; Ken himself didn't know. 
But they all knew the possibilities — ^he would either join the startup fiill-time, join part-time, or leave 
— and therefore dealt with this known-unknown by crafting a buyout agreement that carefiilly laid out 
the rules and price for buying back founder shares should a founder cease to participate in the startup. 
The agreement specifically mentioned the possibility that Ken would not be working for the startup: 
"In the event Burows is not a full-time employee of the Company by April 19, 2000, the other 
Founding Shareholders shall have the right ... to purchase fifty percent (50%) of the Shares owned 
by Burows." 

But in the high-uncertainty world of startups, an awful lot can fall into the category of unknown- 
unknowns. The first step in dealing with them is to work hard to identify as many as possible and 
discuss how things should change if such an event occurs; that is, turn the unknown-unknowns into 
known-unknowns. For instance, no one could have predicted that Microsoft cofounder Paul Allen 
would be diagnosed with Hodgkin's lymphoma, but any founding team can plan for the possibility that 
one of its members is suddenly forced to cut back or leave for medical or personal reasons. Teams 
who are willing to surface and discuss such scenarios will be able to transform some unknown- 
unknowns into known-unknowns and to work out contingent agreements to deal with them Certainly 
the UpDown team, knowing the 50-50 lottery that drove the awarding of work visas, could have 
discussed the repercussions of each possibility and planned how the equity should change if only 
Georg, only Michael, or neither one received a visa. 

No matter how thoroughly teams try to render unknown-unknowns into known-unknowns, there are 
still going to be surprises and upsets. At such times, the team must rely on the trust that has been 
developed among the members. Such trust, if it has had a chance to become strong before tension- 
filled problems arise, allows cofounders to act outside their immediate self-interest — above and 

beyond the call of duty — confident that, in the longer term, they will reap as they have sown.^-^ 
Although it is easy for cofounders to assume that they will always trust each other, the trial by fire of 
founding a startup often burns a team instead of forging a stronger team 

For this reason, the members of a founding team need to take constant care of their mutual trust; at 
any moment, it might be the rope that keeps them from going over the cliff together. Even if a founder 



is trying to take steps to increase the success of the startup, his or her cofounders might see it as a 
violation of mutual trust. Michael's desire at UpDown to have his equity stake match his contributions 
was initially seen by his cofounders as greedy and selfish, rather than as an attempt to make the team 
more stable in the long run. However, his willingness to communicate openly with his cofounders and 
to adjust the equity stakes in the tenplate to meet their needs enhanced the trust within the team. Thus, 
the tension-filled, high-stakes equity-split negotiation is itself a double-edged sword, an experience 
that can leave the team wiser, stronger, and more unified, or can undermine the trust within the team. 
In particular, some founders forget that the equity-split negotiation is one in a series of negotiations 
that will determine the founders' individual and collective success. A founder who pushes hard for 
every last percentage point of equity may leave the other founders feeling they'd better watch their 
backs, which will make it harder for the team to handle the inevitable unknown-unknowns when they 
arise. As Jim Triandiflou observed about such situations, "The juice isn't worth the squeeze." 

The UpDown team's extensive discussions over equity actually solidified their mutual trust. 
Reopening the equity negotiations in February 2007 and arriving at an agreement that they all felt was 
fair gave them valuable insights into each other's motivations and goals, helping them clarify 
Michael's role, appreciate Phuc's contributions, and understand where the startup ranked on Georg's 
list of priorities. Thus, when Georg did not get his visa and left the country after graduation, the team 
was again able to renegotiate calmly. In 2007 and 2008, Georg returned two-thirds of his stock but 
kept one-third in consideration of the time and effort he had contributed to the startup. In contrast, at 
govWorks.com, the extensive negotiations with Chieh destroyed the trust between him and the rest of 
the team 

Protecting Yourself from Your cofounders: Self-imposed Vesting 

Vesting is the most common type of dynamic equity agreement. Vesting terms require founders to earn 
their equity stakes, either over a specified time or when they accomplish specific milestones, rather 
than owning the equity from the start, as is the case with static equity splits. Founders who leave a 
startup before their equity has frilly vested must relinquish the unvested portion to the startup or to 
their cofounders, thus either shifting it to the cofounders who will continue to build the value of the 
startup or allowing them to reallocate it to someone who will replace the drop-out founder. Vesting 
terms thus help serve as "golden handcuffs" that either give each founder financial incentives to 
continue confributing to the startup, rather than dropping out while keeping the frill equity stake, or 
help protect the remaining founders when one founder leaves. Founder-imposed vesting also enables 
core founders to test whether their potential cofounders intend to stay with the startup for the long 
term and to set expectations about involvement and roles. If Robin Chase had proposed vesting to her 
cofounder, she might have learned from the reaction whether Antje planned to join the startup frill- 
time. 

Nevertheless, founding teams usually resist adopting vesting until forced to do so by their first 
outside investors. Founders ofren worry that proposing vesting terms will be seen as a lack of trust in 
their cofounders' dedication, or resent the idea that their own equity will have to vest, or even fear 
how they might be freated down the road before their equity has vested — without considering how 
vesting protects them^^ 

The two major types of vesting are time-based vesting and milestone-based vesting. With time- 
based vesting, each founder who is actively involved in the startup earns predetermined portions of 



his or her equity stake as each month, quarter, or year passes.* This type of vesting assumes that the 
passage of time approximates the addition of value to the startup, a valid assunption as long as work 
proceeds according to plan. But when the work proceeds more slowly than planned, as is often the 
case, founders can earn ftill equity stakes well before making the contributions that had been expected 
of them. In effect, time-based vesting that is too short releases the "golden handcuffs" and increases 
the risk of losing a founder while the startup is still being built. (For this reason, time-based vesting 
has been derided — ^perhaps unfairly — as "paying for a pulse.") 

Milestone-based vesting is one solution to such a problem, but it can cause its own problems. 
Team members earn a specific amount of equity for each of a well-defined set of milestones that, if 
accomplished, would add concrete value to the startup. For business-oriented founders, the 
milestones may pertain to fiind-raising, customer acquisition, revenues, or the establishment of 
partnership agreements. For technical founders, milestones may be tied to completing a prototype, 
conducting a successfiil beta test, or introducing the ftill initial version. 

While this approach aligns each additional award of equity with the addition of value to the 
startup, it is effective only when the team can (a) define objectively when each milestone has been 
achieved and (b) clearly link the achievement of each milestone to the founder(s) responsible for 
achieving it. If the milestones are more subjective, milestone-based vesting can increase the tension 
and conflict as the founders disagree over which milestones have been achieved. If the achievement 
of a milestone is dependent on the efforts of several founders, only one of whom will receive an 
additional equity stake when it is achieved, the others will be motivated to focus only on those 
milestones that affect their own equity stakes, increasing the tension within the team and reducing its 
ability to achieve milestones. 

In fast-changing companies, adopting rigid milestones can be hazardous and should be done 
carefiilly. Down the road, necessary changes in the startup's strategy may render obsolete a milestone 
on which part of a founder's equity depends, requiring another round of tension-filled negotiation 
over equity stakes and milestones or else leaving that particular founder feeling cheated or creating 
destructive rigidity as he or she continues to pursue an obsolete milestone. One founder-CEO related 
how one of his fimd-raising milestones had defined three metrics that had to be met by his next round 
of financing: a certain minimum amount of capital raised, at a certain minimum valuation, before a 
certain date. Even though "I ended up getting multiple VC offers," some of the offers "did not meet the 
criteria [but] were arguably better for the company." This financing milestone, which had been 
designed to align the founder's incentives with those of the business, thus caused a significant 
misalignment instead. As with all high-powered incentives, vesting terms should be designed with 
such unintended consequences in mind. Boards and founders should "stress tesf ' their tentative 
milestones against a variety of scenarios, assessing whether each milestone would work well in all 
the scenarios and adjusting milestones that, in some scenarios, might cause misalignment. 

However, if used wisely, dynamic terms can provide a team with considerable fiexibility in 
dealing with unique situations. For instance, at the corporate-concierge startup Circles, the two 
cofounders focused on the known-unknown of what might happen if either cofounder gave birth to a 
baby. Even though they thought that the new mother would probably want to continue working, they 
used vesting terms to account for the possibility that she might want to cut back on the days worked 
per week. The founders agreed that if that cofounder wanted to cut back to 80% or 60% time, her 
vesting would adjust at that rate. They also agreed that working less than 60%) time would be 
equivalent to leaving the company or would require a more extensive discussion of what to do then. 
At Ockham Technologies, the founders designed a buy-back system whereby if Ken, the idea person. 



cut back to half time, his equity would be cut accordingly, and, if he had not joined by the end of the 
first year, his equity could be conpletely bought out. Conversely, as we saw earlier, the Lynx team 
agreed that the non-idea founder would vest while the founder who had had the idea would not, a 
flexible arrangement that fit that team's needs. 

High Attraction, High Motivation 

We can now better appreciate the power of dynamic agreements. We saw how splitting equity early 
can help attract cofounders but can undermine the founders' motivation and foster fi"ee riding, while 
late splits can bolster the founders' motivation but interfere with attracting cofounders. Dynamic 
agreements offer a way to avoid this trade-off". Defining the equity split (both its core terms and its 
contingent terms) early in the life of the startup allows the core founders to attract cofounders with 
concrete equity terms. Making the split dynamic — changeable in accordance with each founder's 
continuing contributions — whelps keep cofounders motivated to build the startup's value. 

Ockham's early equity split, for exanple, allowed Mike and Jim to be fijlly attracted and 
committed to the startup (by receiving equity early), while its equity-buyback clause encouraged them 
to continue to work diligently for success because they would have to relinquish their shares if they 
"ceased to perform" This contingency plan prevented a possible disintegration when Ken decided to 
drop out. 

CASH COMPENSATION 

For many founders, the main financial motivation to join a founding team is the equity stake rather 
than the cash condensation. Cash-poor startups generally cannot pay much salary or bonus, but 
confident, passionate founders usually believe that their equity stakes will eventually repay that 
sacrifice. For Ockham cofounder-CEO Jim Triandiflou, not receiving a salary was an entrepreneur's 
badge of honor: "April 19 was the first day of Ockham because that's the first day we did not have a 
paycheck. You're not getting paid, so you're now an entrepreneur." 

On the one hand, my analyses show little association between cash and equity within most founding 
teams, suggesting that decisions about the mix of rewards are relatively ad hoc. On the other hand, as 
we will see, some founders have a clear picture of the role of cash compensation among the rewards 
they seek. For instance, within the UpDown team, Phuc Truong had a clearly stated tradeoff he was 
willing to make between cash and equity, trading off less of one for more of the other. 

Two striking patterns of founder compensation are the equality of pay within many founding teams 
and the below-market salaries ofl;enpaid to founders (the "founder discount"), the latter of which can 
give us some insights into founders' attachment to their startups. 

Salary Equality versus Inequality 

While Jim Triandiflou gloried in his initial lack of salary, cash compensation remains important for 
some founders, particularly those accustomed to making a considerable amount. Phuc Truong, for 
instance, was making $200,000 a year on soflware-development consulting contracts and was 
adamant that UpDown compensate him for most of this lost income should he join full-time. He 
wanted a salary of $110,000, and his lower limit was $70,000, but for every $10,000 below 



$110,000, he wanted an additional 0.25% of equity. In contrast, Michael, who was a full-time 
student, cared very little about receiving a salary. Thus, the team ended up with cofounders making 
roughly equal contributions but earning very different salaries. 

Given the differences among founders, we shouldn't be surprised to find the members of a founding 
team earning very different pay. And indeed, some suggestive quantitative analyses of my 2009 
dataset revealed that, of the startups that had not yet raised any outside capital but still had at least 
two cofounders working in the startup, 63% paid the founders unequal salaries.* As would be 
expected, the founder-CEO was often the highest-paid cofounder. However, in almost one-third of 
these unequal- salary startups, a non-CEO cofounder actually earned more than the founder-CEO, a 
striking figure given that, in large companies, the CEO almost always makes significantly more than 

anyone else.^^ 

While salary equality for founding teams is not the rule, it is common in young startups and may last 
for a long time; Lynx's three cofounders received the same salaries throughout the life of the startup. 
In my quantitative dataset, 37%) of the pre-fmancing startups that still had at least two cofounders 
were paying their founders the same salary. This percentage dropped with each successive round of 
outside financing, with only 19% of the teams having salary equality after they had raised their third 
round of outside financing. Salary equality was more common in two-founder teams than in three- 
founder teams — 40% and 33%), respectively. Although the percentages dropped by the third round of 
financing, the relationship between two-founder and three-founder teams remained: 20%) of the two- 
founder teams and 13%o of three-founder teams had salary equality. 

The Founder Discount — Dedication or Exploitation? 

The founders of Lynx Solutions were very discouraged with their cash compensation. At first, they 
had voluntarily accepted below-market — in fact, zero — salaries, believing that it would be better for 
the startup if they used all resources to get it off the ground. When they raised money from venture 
investors, they negotiated with their new board for salaries of $60,000 each. This was still far below 
market, but the founders believed they were "doing the company a favor" by not asking for too much. 
However, when the founders later sought a raise to bring their salaries more in line with industry 
norms (well over $100,000), their board would grant them only $90,000. The three founders were on 
the verge of threatening to leave if the board did not agree to pay them fairly — after all, they 
themselves had responded to such threats from the startup's non- founding employees. As we will see, 
though, such threats were far less effective coming from founders. 

The Lynx founders are not alone in being underpaid. In analyses of my quantitative dataset, I 
examined the compensation of 1,238 founding and nonfounding C-level executives in 528 private 
startups. Controlling for differences across the startups (such as company age, size, resources, and 
industry segment) and differences across the executives (such as equity holdings, educational and 
work backgrounds, and tenure in the organization), I found that founders received $25,000 less in 
salary than equivalent non- founders did — clear evidence of a "founder discount."* 

Two influential theories of corporate management — agency theory and stewardship theory — can 
shed light on the compensation differences between founders and nonfounders. According to agency 
theory, executives are "agents" acting in their own self-interest as opposed to the interests of the 
company; company owners therefore have to structure compensation (and monitor their agents' 
behavior) so as to tie an executive's self-interest as closely as possible to the organization's goals. 



Nonfounding hires are classical "agents," self-interested individuals who take a job for the personal 
income and benefits. Founders need to align this self-interest with the interests of the company by 
structuring compensation to make it contingent on firm performance. In contrast, founders (especially 
founder-CEOs) tend to act more like the principals or "stewards" of stewardship theory — ^people 
who identify closely with their organizations and thus derive higher satisfaction from promoting 
organizational interests than from purely self-serving behavior.* In beginning a startup, founders 
incorporate their startups into their self-identities to the point where their own self-esteem becomes 
deeply tied to the startup's success or failure. While this situation sounds ideal in that it aligns the 
interests of the startup with those of its founders, it also has a dark side: It can severely limit a 
founder's bargaining power for compensation, resulting in the founder discount. 

We thus have two explanations for the founder discount, one voluntary and one involuntary. 
Founders receive a great deal of psychic income and personal satisfaction from working on their 
ideas and may even embrace taking less compensation in order to help the startup. This was why the 
Lynx founders were underpaid — at first. But a labor of love can become a trap after the startup 
receives outside funding or forms a board of directors. Founders like those at Lynx, who feel 
underpaid but cannot bring themselves to leave their beloved startups in order to earn better pay, can 
be forced by their boards to continue to accept a salary discount. 

In the end. Lynx's founders couldn't credibly threaten to leave on account of a raise. As founder 
James Milmo explained, "The board knew we were so attached to the company that we wouldn't 
walk away — we had too much vested in building the company to leave over short-term cash 
compensation. So they knew it was a bogus threat. . . . [Our artificially low salaries] ended up being 
an albatross around our necks" and a significant distraction. 

However, a founder's power to negotiate a raise should increase over time. As a startup grows and 
its founders are forced to cede more and more ownership to investors and employees, we might 
expect them to become less profoundly attached — less like stewards and more like agents. Boards 
would then do well to adjust by closing the condensation gap between founders and nonfounding 
executives; that is, relying less on attachment/stewardship and more on conpensation/agency. This is, 
in effect, what happened at Lynx. When the startup was several years old, the founders finally were 
able to negotiate a bigger slice of the rewards through an equity carve-out. My empirical results 
reinforce the notion that the founder discount shrinks as the startup grows. By the time startups grow 
to 100 employees, there is no significant difference between founder and nonfounder compensation. 
This evaporation of the founder discount is consistent with the disappearance of both its voluntary 
and involuntary causes. 

CLOSING REMARKS 

Splitting the equity is one of the most complicated and tension-filled of the founder's dilemmas. The 
natural inclination is to bypass the tension by taking the sinplest route — an equal and static split. 
Founders should resist this urge, which is fraught with longer-term peril. Instead, they should seek the 
right time to split and use the right approach to doing so. 

Splitting the equity too early is a recipe for continual renegotiation. If the business is still 
amorphous and the team conposition is in fiux, the founders should hold off negotiating the equity 
split. At the same time, an external event (such as an investment offer) or the need to provide a key 
cofounder with clarity about equity stakes (e.g., if he or she is being offered another opportunity) may 



force the team to split earlier than they want. In such a case, the dynamic element described below 
takes on even greater importance. 

When things start to solidify, the founders should engage in a detailed attenpt to match the initial 
split to each founder's expected long-term contributions. The tenplate used by the UpDown team in 
Figure 6.6 offers a model. Although its details do not apply to every founding team (and some of its 
details were controversial even within the UpDown team itself), its overall structure and the process 
that that team followed can be generalized. Adapting this template to a particular startup provides the 
founding team with a structured approach to discussing complex emotional issues and arriving at an 
initial split. As the founders begin to agree on particular terms, they should document those 
agreements in writing to avoid later misunderstandings or miscommunication. Investor Jeff Bussgang 
observes, "I find that the most common source of tension is the 'he said/she said' disagreements down 
the road, which can be avoided by having well-documented, clear agreements or contracts." 

In the midst of negotiating such a tension- filled issue, each founder should remember that working 
together is a long-term proposition and that the equity-split negotiation can make or break the team's 
relationship. Founders who treat the negotiation as a "transaction" and try to maximize their own 
short-term deals may poison the relationship and lose in the long term, securing a larger slice of a 
smaller pie or increasing the chances that there will be no pie at all. 

It is possible that careful negotiation will arrive at the same equal split that a quick handshake 
could have secured. Yet when the time comes for the first round of financing, quick handshake splits 
are associated with lower valuations than negotiated splits (even negotiated equal splits). Why? One 
reason is that the equity- split negotiation often acts as a trial by fire: If the founding team survives the 
negotiation, it is often a stronger team able to tackle tough issues. Better to find out sooner that the 
team has insurmountable difficulties making tough decisions together than to find that out after each 
founder has sunk a lot of time and money into a doomed startup. 

The initial split should be accompanied by a dynamic element. To take an extreme counterexample, 
if a team is sure that (a) its strategic focus, business model, customer base, and competitive situation 
will not change and (b) each founder will remain fully committed and contributing at a high level and 
that nothing will happen in his or her personal life to affect that (which is obviously impossible to 
know), then the team can decide on an equity split once and for all. Otherwise, setting an early equity 
split in stone, without any way to adjust it to expected or unexpected events, is one of the biggest 
mistakes a founding team can make. In a fast-changing startup, early plans often change dramatically, 
making one cofounder's contribution more valuable than expected and another's less so. Even the 
most well-intentioned cofounder can cease contributing as he or she learns about the demands of a 
startup or as life intervenes. Such changes can leave the most important founders undercompensated 
while underperforming founders are overcompensated, leading to destructive tensions within the 
team. 

Rather than succumbing to their natural optimism, founders should structure their equity split on the 
assumption that some aspects of the startup (such as its business model or strategy) and their founding 
team (such as the founders' roles or levels of commitment) will change. They should (a) define how 
foreseeable scenarios should affect the equity split and (b) "plan for the unforeseeable" by including 
buyout terms or similar means by which an underperforming cofounder's equity can be reclaimed by 
the other founders. Such terms encourage each founder to continue contributing and, failing that, let the 
remaining cofounders redeploy the under-performing cofounder's equity in order to replace him or 
her. Each founder is protected and so is the startup overall. 



There is an inherent conflict in the fact that founders' contributions (relative or absolute) can never 
be precisely defined or measured while their equity stakes and cash condensation can be specified 
down to the decimal point. Nevertheless, cofounders have a very strong sense of what is fair or 
unfair. Indeed, my analyses with Matt Marx showed that cofounders who received less equity than 
they appeared to us to deserve (based on their backgrounds and early contributions to the startup) 
were significantly more likely to depart during the startup's early years.* 

Founding teams who want to avoid the potentially disastrous consequences of an early and static 
equity split should do their best to devise a compensation plan (including the equity split) that (a) 
reflects each member's past and expected contributions as accurately as possible and (b) motivates 
each cofounder without seeming unfair to the others. Teams should also keep in mind that the deal is 
never completely done. Circumstances will change, and the equity split and compensation may need 
to change, too, in order to accomplish what these essential tools are meant to accomplish — an issue 
we will examine in Chapter 7. 



* As described earlier in the chapter, quantitative analyses of my dataset revealed an "idea premium" of 10 to 15 percentage points of 
extra equity. 

^ As described earlier in the chapter, quantitative analyses of my dataset revealed a "serial premium" of 7 to 9 percentage points of 
extra equity. 

i As described in Chapter 7, quantitative analyses of my dataset revealed a "CEO premium" of 14 to 20 percentage points of extra 
equity. 

* As described in detail below, "vesting" refers to the requirement that a person earn his or her equity stake over time or by achieving 
certain milestones. It can, and usually should, play a central role in equity incentives. 

* Founders usually cannot split any later than the closing of the initial round of outside financing. 

* A free rider is a person who benefits from a collective effort while contributing little or nothing to it. In this context, it refers to a 
founder who owns equity but contributes less than his or her share of the value creation, at least in part because this lack of contribution 
will not jeopardize his or her equity stake. 

^ By not offering any equity to David and to other people who contributed to developing the business plan but were no longer involved 
in the company, the Smartix team also took a risk that those early "forgotten founders" would later assert ownership rights. For more on 
the potentially severe problems that can be caused by forgotten founders, see Bagley et aL (2003). 

* There are also tax consequences to waiting to split the equity, including having a higher "basis" for the stock and a later start for the 
long-term capital gains tax clock. For more details, see Wasserman et al. (2009). 

* Frank ended up with 500,000 more shares than he would have had he accepted John's initial offer; each share was worth $0.50 ($10 
million valuation/20 million outstanding shares). 

* The macroeconomic environment may also be an influence, though in uncertain and maybe conflicting ways. For instance, in boom 
times, founders may be more likely than in downturns to fight for each percentage point of equity because they believe that an IPO or 
other successful exit is likely. During downturns, on the other hand, when there is no rising tide to raise all boats, founders may be more 
likely than in boom times to fight for every last percentage of whatever equity they manage to create. Future research could examine 
whether either of these — or other — conflicting effects predominates. 

^ Founding teams are even more Hkefy to split equally in small businesses, where 70% of teams do so (Ruef, 2009). 

* That is, the members of a two-founder team each receive 1/2 of the equity, the members of a three-founder team each receive 1/3 
of the equity, and so on. 

* In fact, even in 60%-40% splits, the holder of the larger stake will be diluted to below 50% in the typical financing round. 

* For instance, when handling intellectual property issues, founders must take into account Section 351 of the Internal Revenue Code, 
and need to make timefy and valid Section 83(b) elections if they are adopting vesting as part of the equity split. For more details, see 
Wasserman et al. (2009). 

* Gates and Allen apparently had no mechanism for adjusting their equity holdings in accordance with such events, despite the 
asymmetry in the founders' contributions to pre-IPO Microsoft. They initially split the equity 64%-36% in Gates's favor; Gates's stake 
was 1.8 times that of Allen's. Even though Allen had been absent for the three crucial years before Microsoft went public, examination 
of Microsoft's S-1 prospectus shows that, at the time of Microsoft's IPO, Gates's equity stake was still 1.8 times that of Allen's stake 



(45%-25%). Allen's recent book (Allen 2011) describes Gates's extreme plan to readjust their equity split, a plan that was foiled by 
Allen. It is hard for the heart to bleed for Bill Gates, who despite having gotten the short end of the stick ended up becoming the world's 
richest man, but all founders should learn the importance of incorporating dynamic elements into equity splits, as we will discuss below. 

* The most common time -based vesting schedules use an initial "cliff of a year, at the end of which the founder earns the equity for 
that year, followed by monthly vesting thereafter For four-year vesting, for example, at the end of the first year the founder would vest 
25% of the equity, followed by three years of monthly vesting, with 1/36 of the remaining equity vested each month. 

* Within these unequal-salary startups, the median gap between the highest-paid and lowest-paid cofounders was 30%. (This 
percentage did not vary much by financing rounds, always staying within a narrow range of 29%-32%.) 

* See Wasserman (2006b) for more details on the theories, samples, and results. 

* Stewardship theory tries to identify the psychological and situational circumstances in which an executive will be most likely to 
pursue organizational interests, even if they conflict with his or her self-interest (Davis et al., 1997). Agency problems are caused by the 
separation of ownership and control However, young entrepreneurial firms are a classic instance of the union of ownership and control 
(Fama et al., 1983), making founders more likely to act as stewards or principals than as agents. 

* For more details, including how we estimated how much equity each cofounder "deserved," see Wasserman et al. (2008). 



CHAPTER SEVEN 

THE THREE Rs SYSTEM: ALIGNMENT AND 
EQUILIBRIUM 



In earlier chapters, we saw how Evan Williams's individual decisions about relationships, roles, 
and rewards affected Blogger. Now we can step back to see how those decisions acted in concert 
with each other to cause the Blogger founding team to disintegrate. For exanple, Evan's decisions to 
found with Meg, a prior girlfriend, and share the decision making with her gave Blogger an 
egalitarian structure that was seriously at odds with Evan's motivation — to maintain control of 
Blogger — as seen by his insistence on the CEO title and having more equity than Meg. Meg's status as 
a close friend coupled with the daily sharing of decisions led her to believe she shared control over 
Blogger; eventually Meg refiised to accept Evan as the primary decision maker and left the startup. 
For Evan, deciding instead to found with a stranger or with a junior prior coworker (or to be a solo 
founder) and being carefiil not to share decision making would have aligned the roles and 
relationships with the unequal equity split and his own motivation, and would probably have avoided 
the painfiil and distracting team breakup. Misalignment can cause disaster. 

BEYOND THE EVDIVroUAL DECISIONS: CONTEXT AND LINKAGES 

On a test, each answer you give is right or wrong independently of the others. But for founding teams, 
multiple decisions must be aligned in order to achieve successM outcomes for the startup. A decision 
made at one point can make a subsequent decision turn out badly, even if that subsequent decision 
might have turned out well in different circumstances. Decisions about relationships, roles, and 
rewards not only send startups down certain paths of development themselves, but also work against 
or in concert with one another in the struggle for growth, the acquisition of resources, and the 
retention of control of the startup. As suggested in Figure 7.1 and detailed below, our Three Rs are 
linked. Thus, a founding team's solutions to its relationship, role, and reward dilemmas must make 
sense not only individually but also collectively. As we will see throughout this chapter, the 
alternative to decisions that are aligned is a set of decisions that are at cross-purposes. At best, 
misalignment of the Three Rs will bring tension and dissension; at worst, it can blow up the founding 
team. 



Relationships 

Friends/family vs. acquaintances vs. past coworkers 
Homogeneous vs. diverse teams 




Roles Rewards 
Overlapping roles vs. division of labor Equal vs. unequal equity splits 

Egalitarian vs. hierarchical decision making Static vs. dynamic agreements 

Figure 7.1. The Three Rs System 

There is no perfect alignment that fits all situations and avoids all problems. The founding of 
Ockham Technologies, for example, began with a relationship decision: Jim chose to found with 
Mike, his former subordinate, and Mike chose to found with his former boss. They then chose roles 
and rewards that were aligned with their previous relationship: Jim became CEO with a 50% equity 
stake and Mike became vice president with a 30% equity stake. This set of aligned decisions worked 
well. But it would surely not have worked so well at UpDown. Within that team of strangers and new 
classmates, crowning a CEO and adopting a very unequal equity split at the time of founding would 
have been out of alignment with their decision to found with people about whose capabilities and 
motivations they still knew so little. (Below, I describe theoretical reasoning and empirical evidence 
that support the riskiness of having these factors misaligned.) For this team, adopting an egalitarian 
approach to decision making, assigning overlapping roles to the two similar cofounders, and splitting 
the equity equally except for a small idea premium was a better-aligned system of decisions — yet one 
that would not have been aligned for Jim and Mike at Ockham. 

Several times in Chapters 3 through 6, we saw the unhappy consequences of making quick and easy 
Three R decisions by default. The potential for disastrous consequences as a result of decisions like 
these is multiplied by the fact that there are linkages between the Three Rs. Understanding the causes 
of tension can be difficult for teams: Even one seemingly small decision, if misaligned with others, 
can tip the balance, and even one founder who is out of alignment with the others can cause an 
emotionally draining breakup. Because the effects of seemingly unrelated Three R decisions can be 
cumulative and hard to pinpoint, the concept of alignment offers an important tool for troubleshooting 
and resolving tension within a founding team As we will see, the Three Rs framework gives such 
teams a structured approach for diagnosing the causes of tension — that is, for understanding the nature 
of the misalignments, resolving them, and creating a new equilibrium 

We have also seen that almost all the decisions a founding team makes are trade-offs. Even the 
"besf ' decisions about relationships, roles, and rewards bring risks as well as benefits. When these 
Three R decisions are aligned, those risks and benefits are brought into productive equilibrium But, 
even when a team reaches a consistent equilibrium across the Three R system of decisions, it must be 
careful to reevaluate the system as the startup evolves and as unexpected events occur. As UpDown 
evolved, for example, the team's Three Rs decisions — so well aligned at first — later became wholly 



inconsistent, bringing the team to the brink of a breakup, until it could achieve a new equilibrium The 
need for dynamic alignment will be addressed later in this chapter. 

We will start by examining the linkages between each pair of Rs — relationships-roles, 
relationships-rewards, and roles-rewards — to see how they can be aligned or misaligned, and then 
examine the need for a dynamic approach to the Three Rs system 

LINKING RELATIONSHIPS AND ROLES 

In this section, we will look more closely at the linkages between relationships and roles. For 
example, how does the prior relationship influence the roles that founders play within their startups? 
Are there optimal ways to delegate roles and decision making depending on the type of prior 
relationship? We will treat each of our three types of prior relationships in turn, examining the role 
implications of founding with friends or close family, coworkers, and strangers. 

Founding with Friends and Family 

Prior relationships imply expectations of two kinds. We have descriptive expectations based on what 
we know from experience about another person's abilities, feelings, behavior, and so on. We also 
have prescriptive expectations that a friend or relative will not hurt our feelings, will not double- 
deal, will not throw away a friendship or family unity for the sake of a business advantage. But if 
these expectations are founded only on nonwork experience, they can be severely tested in the context 
of founding a startup, which can often be an extreme form of work experience. Such was certainly the 
case when Steve Wozniak found himself on the short end of a dishonest business deal made with the 
person he was expecting to behave as his best buddy. 

Aside from the parent-child relationship, we do not expect our relatives and friends to have any 
more power over us than we grant them out of respect or affection. It would seem, then, that an 
egalitarian role structure would be better aligned with the decision to found with social relations than 
a hierarchical role structure would be. And indeed, I found in my dataset that the closer the prior 
social relationship, the more likely the team is to adopt an egalitarian approach to decision making. 
Teams of friends, for example, were 8% less likely to name a CEO than were teams of nonfriends. 
Used to being peers on a social basis, such cofounders tend to transfer the peer relationship to their 
professional relationship within the startup. The Stanford Project on Emerging Conpanies (SPEC) 
also found evidence of this strong linkage within high-tech startups. Of all startups in their study, 30% 
adopted the "commitment" model in which (a) employees are selected and monitored based on 
cultural fit and are attached to the organization like a family and (b) the founders emphasize universal 
buy-in for all major decisions. But of the startups in which family or friends of the founder were 
listed as key partners, fiilly five-sixths adopted the commitment model. ^ This effect of founding with 
family or close friends provides evidence that prior relationships can have a substantial influence on 
decision-making structures. 

Founding with Coworkers 

In contrast to teams with prior social relationships, teams of prior coworkers often find it more 
natural and better aligned to adopt a hierarchical structure, especially if one cofounder is more 



experienced than the other or if one has worked for the other. The roles and relationships reinforce 
each other rather than operate at cross-purposes, and there is less tension than there would have been 
otherwise. As noted in Chapter 4, this was the decision made by the founders of Ockham 
Technologies; their subsequent tension underscores the initial alignment of their original decision. At 
first, these prior coworkers — Mike had worked for Jim — adopted clearly hierarchical roles: Jim was 
founder-CEO and Mike was a vice president reporting to him. Tension arose when Jim included 
Mike in most of the decision making and on the board of directors. This de facto egalitarian decision- 
making structure gave Mike a role in his own mind that was at odds with his official role and, more to 
the point, with the role he had in Jim's mind. Thus, their relationship and their roles were no longer 
well aligned and had to be brought back into alignment. One possibility was to change the 
relationship: That is, they could agree to think of each other as peers rather than as former boss and 
subordinate — a hard shift to make. The other possibility was to change their roles. That is, Jim could 
take care to work with Mike as a vice president and not as a co-CEO. In fact, they took the latter route 
to realignment. 

The cofounders of StrongMail — Frank Addante and his former subordinate TimMcQuillen — faced 
a similar dilemma. They, too, adopted a hierarchical role structure seemingly well aligned with their 
earlier work relationship: Frank became CEO and McQuillen was second-in-command. They, too, 
found this arrangement somewhat at odds with the more egalitarian day-to-day experience of founding 
a startup together. "The hardest part," Frank explained, "was transitioning from a role where he 
worked 'for' me to a role where we were building a business together. But [at Strong- Mail], I was 
still the CEO so technically he worked 'for' me. . . . [Ojverall, it was a big positive because we had 
already developed mutual [professional] trust for each other." 

In each case, the two cofounders discovered that early on, the demands of the startup made it more 
natural to adopt overlapping roles and collective decision making. This required an adjustment from 
their prior hierarchical relationships. However, as the startup grew and required a hierarchical 
structure, the cofounders' hierarchical roots enabled them to make that transition more smoothly. 

Founding with Strangers 

Relationships and roles must be aligned in terms of division of labor as well as of hierarchy. 
Choosing a cofounder with whom one has no prior relationship can affect the degree to which roles 
within the team are well defined and exclusive, though the effect can vary from one team to another. In 
part, this is because there is an inherent difficulty in aligning roles (or rewards) with what is actually 
the lack of a relationship. 

One approach that at first seems to make sense is to adopt a clear division of labor in order to 
circumvent uncomfortable ambiguity. As Tim Westergren, founder of Pandora Radio, explained, 
"You need history if you want to do overlapping roles; if you don't have history, it's probably better 
to go for more defined roles. If you have overlapping roles [but no history together], you start to look 
at each other and wonder if you both need to be here." Such early clarity is understandably tempting 
to entrepreneurs as a way to reduce tension in the short term 

However, a quick division of labor carries the longer-term risk that, as the founders get to know 
each other and adapt to life within the startup, one or more of them will see a reason to reallocate 
roles. As Vivek Khuller of Smartix observed, "When the team doesn't know each other very well, 
where there are different domains, where you have little history of working together, it's best to delay 



it because things are still unknown and changing." Vivek's approach — adopting vaguely defined and 
overlapping roles until the cofounders know enough about each other to apportion the labor more 
precisely — can help avoid the risks of allocating roles blindly, but at the cost of early tension over 
roles. For instance, it took two earlier startups for members of FeedBurner's founding team, Dick 
Costolo and Steve Olechowski, to finally divide the roles among themselves. Even though they had 
worked together in the past, Dick did not realize that Steve would want to move away fi^om pure 
software engineering and into business-oriented tasks — things that had traditionally been Dick's role 
— causing tension that was resolved as their startup grew enough to accommodate two business- 
oriented founders. Dick reflected, "There are lots of things you have to learn about each other. You 
can't anticipate how the executive team should split things up or what roles you should each play." 
But by the time they started FeedBurner, Dick explained, "we didn't need to discuss roles or 
equity. ... We knew that Steve would be doing much less on architecture and programming. Now it 
was a given that he'd be the business-operations guy, dealing with legal, accounting, operations." 

Reflecting on his team's experiences across their four startups together, Dick Costolo said, 
"Mapping out our roles early would certainly have helped." Amid the chaos of a startup, such role 
clarity is particularly attractive, and Dick naturally would prefer to have avoided the tension he 
experienced with Steve. However, with teams who are just getting to know each other, jumping too 
soon to deflne roles can be a big mistake, for if the roles are not allocated well and then have to be 
reassigned, tensions will be heightened rather than reduced, and the quality of work produced by the 
reassigned founder is likely to suffer. Thus, had Dick gotten his wish and split roles early with 
cofounders he barely knew, the team might not have made it through multiple startups together and not 
been together to build FeedBurner into the success it became. 

LINKING RELATIONSHIPS AND REWARDS 

Prior relationships have powerful determining effects, not only on role decisions, but also on the oft- 
contentious decision of how to split equity in a startup. A founding team's reward decisions must be 
aligned with the cofounders' prior relationships. As we have already seen in Chapter 6, when the 
differences within a team are below a certain threshold, the team tends to split the equity equally, but 
when they're above that threshold, the team tends to split the equity unequally. However, the team's 
prior relationships can have a strong impact on how high or low that threshold is. In particular, my 
analyses with Thomas Hellmann show that family relationships tend to raise the threshold. All other 
things being equal, if a founding team included at least one set of family members, it was more likely 
to split the equity equally. Researcher Martin Ruef's analyses of small-business equity splits also 
show that family relationships can skew equity stakes: When a core founder included a family 
member on the founding team, that family member received an equity stake that was 1.11 times the 
stake of a comparable non- family cofounder.^ This can make for a serious misalignment if the family 
and non-family cofounders are not actually making roughly equal contributions to the startup. The 
cofounders of Dimon, a computer-game startup that Georg Ludviksson had cofounded before 
cofounding UpDown, had exactly this problem Three of the six cofounders were siblings and the rest 
were close friends, so an equal equity split was far "easier" than a conflict- filled negotiation among 
friends and family members. But when investors required the three nonfamily cofounders to buy out 
the three family members who were no longer active in the startup, the buyout was "really painful" 
and left the three siblings feeling "unfairly treated." 



We have already cautioned against the impulse to make quick equal splits in an attempt to avoid 
difficult negotiations. Yet in practice, we consistently see the association of family relationships with 
equal equity splits. To explain this tendency and determine its implications for the stability of the 
founding team, let's take a step back to examine equity theory.-^ Equity theory highlights the tight 
linkage between social factors (relationships) and economic factors (rewards). Prima facie, founding 
teams can be categorized as operating under a social logic or a business logic, depending on the type 
of prior relationship shared by the cofounders. For teams operating under a social logic, preserving 
personal relationships takes precedence over maximizing business success; for teams operating under 
a business logic, maximizing business success takes precedence over preserving personal 
relationships. Even a husband- and-wife founding team has to care about growth and profits, but such 
a team (if it is truly operating under a social logic, as one would expect) would choose a suboptimal 
business outcome if the optimal outcome could be achieved only at the cost of their marriage. When it 
comes to splitting equity, teams operating under a social logic will find the rule of equal 
distribution — distributing rewards equally, even if individuals have very different levels of 
contribution"^ — to be most aligned with the priority they give to their relationships. For instance, 
college roommates who operate under a social logic reward each other equally despite differences in 

individual performance.^ In contrast, teams without prior social relationships and operating under a 
business logic (e.g., teams of former coworkers or teams of strangers) will find the rule of equitable 
distribution — distributing rewards in proportion to the value of each individual's contribution — to be 

most aligned with the priority they give to individual performance and business success.^ Equity 
theory ultimately concludes that the best equity split for one type of team could be the worst equity 
split for another type of team, depending on the dominant logic operating in the specific circumstance. 



Basis for 
equity split 



Rule of equal 
dtstrihutiou 



Prior relationship 



Social relationship 
(family, friends) 



Stable team 



Rule of equitahle 
distrihution 



Unstable team; 
inconsistent with 
social logic 



Prior coworkers 



Unstable team; 
inconsistent with 
business logic 



Most stable team 



Figure 7.2. The Linkage between Prior Relationships, Equity Splits, and Team Stability 



Teams whose prior relationships and equity splits are aligned will, on average, be more stable and 
sturdy than those that are not. Indeed, Matt Marx and I, in our examination of founding-team turnover, 

found a compelling linkage among relationships, rewards, and team stability.^ Figure 7.2 integrates 
the results of our quantitative analyses of founding-team stability with the field data from this part of 
the book and the theoretical lens of equity theory. Within the group of teams with prior social 
relationships (i.e., the ones presumably operating under more of a social logic), the most stable teams 
(i.e., the teams that are still frilly intact after various amounts of time) tend to be the ones who split the 
equity equally. In confrast, within the group of teams with prior professional relationships (and 



presumably operating under a business logic), the most stable teams tend to be the ones who split 
equitably. 

The results summarized in the figure drive home the dangers of making a key decision without 
considering its alignment with all of the other key decisions. Without knowing the cofounders' prior 
relationships, we can't state categorically that an equal split is better or worse for team stability than 
an unequal split. Knowing the type of relationship, however, does indeed lead to a particular 
approach to splitting that seems better aligned for team stability. 

For founding teams of former coworkers, equitable (performance/contribution-based) equity 
allocations are well aligned with the cofounders' relationships in another way. Such allocations can 
act, over time, as a sort of natural selection: Poor performers may become dissatisfied with their 
comparatively lower rewards and leave,^ increasing the team's proportion of high performers and 
therefore its long-term performance.^ Such a weeding-out process would be painfully unaligned with 
the relationship choices of a team of relatives or friends. 

Cofounders who have not only worked together previously but also founded startups together 
present a unique set of circumstances. Such "serial founding" teams (e.g., FeedBurner's team) 
possess far more knowledge about each other and a better feel for each other's commitment levels 
and motivations than any other grouping. They thus can make a well-aligned equity split much sooner, 
gaining the advantages of an early split without suffering many of the disadvantages otherwise 
associated with making such a decision too soon. However, even these teams should seriously 
consider adopting dynamic terms within the split, for the unexpected — illnesses, family crises, even 
visa problems — can happen to them, too. 

Cofounders who have no previous acquaintance, either social or professional, need to align 
rewards with the relationship decision they have made; that is, with the lack of a prior relationship. 
The key here is for the cofounders to grasp how frilly they do not know each other. Resumes, online 
bios, and even third-party recommendations can't tell you what makes a person tick and what 
contributions he or she will make, or fail to make, in the unpredictable growth of the startup, nor can 
they tell you a person's commitment level or whether he or she has what it takes to persevere through 
the scariest parts of the enfrepreneurial roller-coaster ride. Such teams have much to gain by holding 
off their equity split until they have learned a lot more about each other, although they may pay a price 
for this benefrt by sacrificing some of their ability to attract key cofounders (as did the Smartix team 
when it waited to split and thus lost the potential cofounder who had deep knowledge of and contacts 
in their industry). Once again, dynamic equity splits are an invaluable way for such teams of 
unfamiliar cofounders to deal with the frade-offs between attraction and motivation when facing the 
challenges of founding a startup. 

LINKING ROLES AND REWARDS 

While it is not surprising that hierarchy and division of labor affect equity splits, it is less commonly 
understood how their absence can also affect equity splits. Overlapping, nondistinct roles make it 
harder for cofounders to determine the value that each has contributed or will confribute to the startup 
and therefore make it more likely that the team will decide to split the equity equally. An unequal split 
would not be well aligned with such uncertainty about who really deserves what share. 

On the flip side, clearly defined hierarchical roles tend to increase the likelihood of an unequal 
equity split. In particular, the person receiving the CEO title is ofren expected to contribute more 



value to the startup than anyone else. (A common exception is founding scientists or technologists in 
startups based on intellectual property they developed.) Evan Williams equated the CEO role with a 
bigger equity stake, as did Frank Addante; it wasn't until Frank was CEO in his later startups that he 
felt he could actively lobby for the largest piece of the pie. Both expected to make the largest 
contribution to the startup's success. At UpDown, Michael received a larger equity share than his 
original equal share and eventually became CEO as it became clear that he was making and would 
continue to make the largest contribution to the startup's success. 

Indeed, my analyses with Thomas Hellmann show that, controlling for the other differences across 
founders, founder-CEOs receive a CEO premium (i.e., an additional amount of equity for being CEO) 
of 14 to 20 percentage points. The premium for founder-CTOs is lower — 5 to 8 percentage points — 
but also highly significant from a statistical perspective. The idea premium discussed in Chapter 6 
may also be, in part, a premium for adding greater value to the startup by initially articulating its 
vision, attracting other cofounders, and completing other critical early tasks often performed by the 
idea person. (Founders who serve on the official board of directors also play a different role than do 
otherwise similar founders. It is hard to assess whether such founders receive di board premium, 
because founders usually serve together on an informal board and often do not form an official board 
until after the equity split has been decided. However, serving on the board does provide executives 
with higher cash compensation than is received by equivalent nondirector executives; eventual 
board membership could also affect founders' equity stakes.) 

It isn't only teams that have to align their role and reward decisions; individual cofounders do, too. 
cofounders who are motivated more by financial rewards than by control considerations should be 
willing to give up a better title or more central role in order to gain additional equity or cash 
condensation, while cofounders for whom becoming CEO is important might have to give up 
financial rewards to their cofounders in order to secure the title they want. My equity- split dataset 
suggests that some founding teams may be taking exactly this course: While in 39% of the startups, the 
founder-CEO held the largest equity stake (recall, there was an average CEO premium of 14-20 
percentage points in this dataset), in 33% of the startups, the founder-CEO's equity stake was on par 
with his or her non-CEO cofounders, and, in 28% of the startups, a non-CEO member of the founding 
team received a higher equity stake than the founder-CEO did.* In some of these latter startups, the 
founders made conscious trade-offs between roles and rewards so that each cofounder would gain his 
or her own highest priority and be motivated to stick with the startup and give it his or her all. In 
others, the non-CEO contributed a scarce and critical nonmanagerial skill. 

More broadly, as described briefiy in each chapter of this part of the book but delved into more 
deeply in Part III, this need to make trade-offs between (a) roles and decision-making control and (b) 
financial rewards recurs throughout each of the key dilemmas faced by founders, beginning with the 
first decisions they make regarding when to found and how to build their founding teams and 
continuing through the hiring, investor, and exit decisions that we will discuss in Part III. Chapter 1 1 
highlights how these recurring decisions help determine whether the founder achieves a Rich 
outcome, a King outcome, neither, or, in rare cases, both. 

The relationship between roles and rewards can also flow in the opposite direction; people who 
receive larger equity stakes may gain influence. One founder admitted that whenever people voiced 
an opinion or disagreed during a discussion, their relative equity stakes came to mind and affected 
how much weight he gave to their opinions. 

The interaction of role decisions and equity decisions can have another inportant, long-term 



economic impact. Equity stakes are a scarce resource, and the allocation of equity is usually more 
efficient when the team has distinct roles than when the founders have overlapping roles and are 
therefore somewhat redundant. Instead of using scarce percentages of equity on overlapping founders, 
the startup could have used that equity to attract a cofounder with a missing type of human, social, or 
financial capital who would make the startup more valuable. In this way, the rewards structure would 
be better aligned with people's roles. Or, further down the line, those equity stakes could have been 
used to attract hires or investors to fill those holes and add value, as will be explored in Part III 
below. (This issue is exacerbated if the equity stakes are being held by fully vested drop-out 
founders, whose equity could have been reclaimed and redeployed if the team had adopted a dynamic 
equity agreement.) Thus, early inefficient uses of equity rewards within the founding team can leave 
the team handcuffed when it needs to fill necessary roles. 

CLOSING REMARKS 

When each pair of decisions is aligned, the team has achieved a "Three Rs equilibrium" in which 
tensions will tend to be low and the team can focus on building the value of the startup. But even then, 
there are at least three challenges that make the equilibrium particularly difficult to maintain. 

Communication 

Relationships, roles, and rewards are all sensitive issues that most teams try to avoid discussing. As 
the founder of a marketing-automation company observed to me about the tensions within his team, 
"The elephants in the room that no one wanted to bring up all fall into those categories." Some types 
of prior relationship, such as previous work experience together, facilitate raising role and rewards 
discussions within the team, but most types of relationship lead teams to avoid the tension of tackling 
those issues. If two cofounders find themselves stepping on each other's toes, for example, or one 
finds himself or herself working much harder on an exciting lead while the other is working on 
incremental product development, they may keep their fi-ustrations to themselves so as not to disturb 
the team spirit. In a classic case of "not discussing elephants," it was the thought of losing his best 
friend that kept Steve Wozniak silent about his increasing unhappiness and frustrations at Apple. 
However, even reaching verbal agreements is often insufficient; founders should capture such 
agreements in writing to ensure clear communication and avoid later disagreements. 

Change 

Startups regularly face unexpected changes in personnel, competition, technology, economic 
conditions, regulation, and so on. Even if a team has achieved early alignment, it must be ready to 
adapt its arrangements or must have done the hard work early on to craft effective dynamic elements 
in its agreements, so that it can realign itself after the unexpected happens. As we will see in Part in, 
a particularly jarring change is when outside investors enter the picture, often changing the roles 
played by the founders, forcing realignment of their equity holdings and altering their relationships. 



Inertia 



Although early alignment can contribute to early success, it can also cause inertia and complacency 
that prevent effective adjustments to rapid shifts in the organizational environment.^^ So even when 
things seem to be proceeding smoothly, teams must proactively look ahead to the changes that will be 
caused by startup growth and evolution, plan for those changes (e.g., by setting expectations about 
when each founder might relinquish his or her C-level position), and regularly reassess the alignment 
of those arrangements. 

Both change and inertia challenged the UpDown team, which seemed to have achieved an early 
equilibrium. The team's equity agreement in November 2006 was a quick, mostly even split, with 5% 
more to Michael for having the idea. The team shared decision-making power but assigned 
reasonably clear roles: Michael and Georg, who had similar backgrounds and skills, would focus on 
developing the business, while Phuc (the software engineer) would work on product development. 
All the founders planned to work fiill-time on the startup starting in the summer of 2007. Thus, the 
fairly equal equity split was in alignment with the expected contributions of the team members. 

However, as the startup evolved over the next two months, Michael took the lead in business 
development, while Georg vacationed in Europe with his family and Phuc continued to work on his 
previous consulting commitments. Michael concluded that Georg' s and Phuc's motivation and 
commitment — and therefore their contribution to the startup — were not equal to his. He recalled his 
frustration with the way the team responded to the first real possibility of gaining an outside investor: 
"I had planned to go on [a] student trip, but I cancelled that to work fiill-time on [UpDown]. At that 
point in time I also talked to Georg, but he said he was going to stay in Iceland until the end of the 
vacation. And Phuc was still involved in his prior consulting projects. So I was feeling very alone . . 
. while I was here working fiill-time. None of the [other] two were really working hard at that point in 
time. ... I [was] kind of scared [about] whether this [was] the right team to work with." The roughly 
equal roles and rewards that had previously been agreed upon were no longer in alignment, and the 
cofounders' relationship had progressed from being relative sfrangers to knowing each other much 
better. 

Michael's crisis brought the team to the brink of dissolution and sparked a renegotiation of their 
roles and rewards in February 2007. The team recognized that their original assessment of each 
founder's past and fiiture contributions had been incorrect, decided to award Michael more equity in 
consideration of his more involved role, and reached a new equilibrium. Yet that summer, an 
unexpected event threw the Three Rs out of alignment again: Georg failed to obtain a work visa in the 
visa lottery while Michael did receive one. Again, the cofounders had to realign their roles and 
rewards, with Michael receiving the CEO title and Georg receiving a greatly reduced share of the 
equity since he could no longer contribute an equal share of the work. This was their third 
equilibrium within a year, but they would have paid a far stiffer price had they refiised to realign their 
relationships, roles, and rewards to their oft-changing circumstances. 

Throughout this part of the book, we have examined how founders' natural inclinations can create 
pitfalls for their decision making and for their ability to build their startups. Passionate, confident 
founders often hear about the problems introduced by mixing social relationships with business roles, 
adopting egalitarian decision-making processes, and agreeing to quick-handshake equity splits, but 
want to believe that they are different, that they will be able to deal with it. One founder lamented to 
me, several months after his startup had failed, "Even after you showed us the data about the risks of 
founding with a friend and of splitting 50/50, we just figured, 'Those things happen to other people, 
they won't happen to us. ' " 



cofounders who make these common decisions should know that they are playing with fire. 
Founder passion and confidence can blind them to the future implications of these early decisions. 
Heightened tension and instability within the founding team can make it hard for the startup to scale 
smoothly and can undermine the founder's ability to grow value while maintaining control. Founders 
are therefore imperiling their startups, and possibly their personal lives, if they do not force 
themselves to plan for the worst and to vigilantly watch for signs that their Three Rs equilibrium has 
fallen out of balance and needs to be reevaluated or renegotiated. 



* In half of the startups in which a non-CEO had the largest equity stake, he or she had received a VP -level title (which is 
unambiguously lower than a CEO title); in almost aU of the others, the founder-CTO or founder-CSO received the largest equity stake. 



PART III 

BEYOND THE FOUNDING TEAM: HIRES AND 
INVESTORS 



INTRODUCTION 



The dilemmas do not end once the founding team is in place. Founders continue to face momentous 
decisions about whether and how to involve other important parties in the startup. In particular, if the 
team has remaining holes, or as growth introduces a need for new human, social, and financial 
capital, the founders have to look to nonfounders who can provide these resources. The two most 
important of those parties are nonfounding hires — people who join the team as employees — and 
investors, who provide capital and might join the board of directors. Any startup that wants to grow 
beyond the founding team has to face the hiring dilemmas of whom to hire, what roles they should 
play, and how to structure their rewards. Although fewer startups face the full range of investor 
dilemmas, most startups face important decisions about sources of financing and will benefit from 
understanding the frill range of options and their repercussions. Ofren, the most important of those 
repercussions is founder-CEO succession, an outcome we examine in depth at the end of this part of 
the book. Figure III. 1 places these new decisions about hires and investors in the context of the 
founding dilemmas examined throughout this book. 



Should I found 
now? 



I 



Yes 



Should I be a 
solo founder? 



No 



No 



Remain nonfounder 



Founding-Team 
Dilemmas: 

• Relationships? 

• Roles? 

• Rewards? 



Yes 



Beyond-the- Team 
Dilemmas: 

• Hires? 

• Investors? 

• Succession? 



Figure III. 1. Beyond-the-Team Dilemmas, in the Context of the Broader Set of Founding Dilemmas 



Founders must grapple with hiring and investor dilemmas over a long period, during which their 
startups can change dramatically regarding resources and the degree of formalization.^ During the 
earliest startup stage, when the business idea is still being developed, the startup has little in the way 
of formal structure or processes and is usually severely resource-consfrained. The founding team is 
often tight-knit, and the culture is informal and emphasizes creativity. Startups then go through a 



transitional stage that is most clearly characterized by two inflection points that affect the degree of 
formalization and the availability of resources: closing on the initial rounds of outside funding and 
conpleting the development of the first product. Both changes add resources (outside capital and 
customer revenues, respectively) but also call for new skills and processes within the organization. 
The division of labor deepens and decision making increasingly shifts away from the founders. As the 
startup enters the mature stage, it becomes relatively well funded and usually has a steady revenue 
stream from a standardized product or service line. Decision making is decentralized across functions 
and becomes more hierarchical; there is much less of a premium on flexibility. New executives, often 
with prior experience in larger companies or with specialized skills, often replace or become 
managers above the early employees (or even founders), who do not have the experience, expertise, 
and networks to lead and grow an entire division or department. 

As the startup progresses through these stages,* founders' preconceived ideas about how they want 
their organization to look and operate sometimes conflict with the needs of the growing organization 
or the agendas of the new players, adding complexity to the founders' hiring and financing decisions. 
Attracting hires and investors forces the founders to make difficult trade-offs, because the more 
scarce and valuable an outside resource is, the more its owner can demand for it.^ (In the case of 
personnel, the owner is the resource.) To attract the best hires, for exanple, founders have to give up 
not only enough cash condensation and equity ownership but also some level of control over 
operational decisions; skilled people usually don't like to be told what to do. To attract the best 
investors, founders have to give up significant amounts both of equity and of control over many 
board-level decisions. As we will see, this means that attracting the best resources can put severe 
constraints on the founders. It can even cost them control of the CEO position and of the board of 
directors. 

OVERVIEW OF CHAPTERS 

The first two chapters of this part of the book introduce the major players and dilemmas in these 
beyond-the-founding-team decisions. The third chapter delves into what is sometimes the ultimate 
consequence of involving outside investors, founder-CEO succession, which requires the hiring of the 
first nonfounding CEO. A brief overview of the chapters in this part follows: 

Chapter 8: Hiring Dilemmas — Even after launching the startup, the team may still have holes 
or the startup's growth may require new people with particular skills. In each stage of 
growth, founders face important trade-offs regarding whom to attract as employees (i.e., 
whether or not to hire people with whom they already have relationships), what roles to give 
them, and what rewards to offer in order to attract and retain them 

Chapter 9: Investor Dilemmas — Investors can provide founders with human capital, social 
capital, and financial capital with which to build the startup. However, different types of 
investors provide very different amounts of value and introduce different risks for the 
founders. In this chapter, we compare the trade-offs of taking money from friends-and- family 
investors, angel investors, and venture capitalists. 

Chapter 10: Failure, Success, and Founder-CEO Succession — The replacement of the 
founder-CEO with the first non- founding CEO is a critical inflection point for many startups. 
We will look at why and how this happens (there are numerous variations), who triggers the 



change and how that affects the succession process, and what are the best and worst practices 
for the transition process. 

Unless otherwise noted, the data in this part of the book come from the Ml decade of surveys, 
2000-2009, which includes more than 19,000 founding and nonfounding executives from more than 
3,600 private startups. 



* Although we outline this general sequence of stages, different functions within startups may go through them at different rates. The 
technical function, for example, may reach Stage 3 (mature) while the sales and marketing function is still in Stage 1 (startup). 



CHAPTER EIGHT 

HIRING DILEMMAS: THE RIGHT HIRES AT 
THE RIGHT TIME 



Either as an alternative to attracting cofounders (see Chapter 3) or as a way to augment a 
founding team that still has gaps, hires play an important role in the growth of many startups. Yet, 
even the most adept founders — those who have proven their leadership abilities by assembling a 
well-tunctioning founding team — can find subsequent hiring decisions difficult and dangerous. 
Founders naturally bring their own preferences to these decisions, but often fail to appreciate how 
growth and change will force them to rethink their preferences and realign their organizations. 
Decisions that seem ideal at first can prove disastrous, as enployees at all levels struggle to scale 
with the maturing startup. 

For a serial founder, decisions that worked well for one startup can be entirely wrong for another. 
When Evan Williams founded Blogger, his vision of universal access to self-expression on the 
Internet was more important to him than maximizing the startup's value. He structured the startup like 
a loyal tight-knit family; his first hires were dedicated fi*iends and colleagues, mainly young self- 
taught programmers like himself Later, to keep Blogger running when money was scarce, Evan relied 
on volunteers, who were willing to work for the "love" of the company and its vision, and contract 
employees recruited fi"om Craigslist and blog posts. He hired young and cheap. 

By the time of his next startup, Odeo, Evan had worked several months at Google (which had 
bought Blogger) and had established an entirely different vision for his new startup: a more 
professional organization that valued the advanced experience and expertise of its employees and that 
would achieve maximum value as quickly as possible in anticipation of a lucrative exit. To this end, 
he raised a large amount of VC ftinding and used it to employ executive recruiters to hire several 
talented programmers and experienced, high-level executives to head up marketing and engineering. 
Evan's hires at Blogger had required close supervision, but now Evan had to relinquish decision- 
making power to these senior hires. 

How can we understand the decisions that Evan made and how they affected his startups? 

The work of the Stanford Project on Emerging Conpanies (SPEC) in its study of enployee- 
relations "blueprints" in high-technology startups provides a useftil framework for understanding 

founders' hiring decisions.^ The study sorted the responses of a large sample of founders along three 
dimensions: recruitment, rewards, and control. Respondents articulated three bases of recruitment, 
three bases of rewards, and four means of coordinating and controlling work.* Although there are 36 
(3x3x4) potential permutations of these factors, five of those permutations — which the researchers 
call "blueprints" — accounted for 67% of the startups."'" SPEC also found that sticking with the chosen 
blueprint was a vital factor in a startup's success. Throughout the course of the study, only 10.9% of 
startups deviated from their original blueprint along all three dimensions, and they were 2.3 times as 
likely to fail as firms that kept to one blueprint.^ 



Viewing Evan Williams's experiences through this lens, we can see that at Blogger, where Evan 
was motivated to pursue a vision, he used a "commitment" blueprint — recruitment was for fit with the 
family-like culture, pay was low because employees were willing to work for love of the startup, and 
control was informal. On the other hand, when he was motivated to maximize Odeo's value, Evan 
used a "bureaucracy" blueprint — recruitment was for experience and tunctional skills, pay was 
commensurately high (requiring outside financing), and control was through a formal hierarchy. 

As we have seen throughout, decisions in one domain can affect another domain. For instance, 
Evan's choices of hiring blueprints affected the rate at which his startups passed through the stages of 
the organizational life cycle. His preference for control over Blogger — his wish to keep it focused on 
his vision — caused him to hire a small number of inexpensive employees, but this also meant that 
Blogger would grow more slowly than it might otherwise have done. With Odeo, Evan was more 
motivated by the prospect of a quick and profitable exit and took a professional — and resource- 
intensive — approach to hiring that would accomplish that. 

The dilemmas that founders face in choosing new hires, like those they face in choosing 
cofounders, fall into the Three Rs framework of relationships (whom to hire), roles (what positions to 
create or upgrade, when to do so, and what types of people to hire into them), and rewards (the 
compensation and equity used to attract and retain hires). Founders have to make different Three Rs 
decisions depending on their motivations and blueprints, and then adjust their decisions during the 
different stages of the startup life cycle. 

RELATIONSHIPS 

Where do startups find their senior executives? Do those hires already have strong ties to the founder- 
CEO or to others in the startup?* As with many of the decisions examined in this chapter, the answers 
depend in part on the stage of the startup. 

The Founder-CEO's Hires versus the Investors' Hires 

Founder-CEOs of fast-growing startups often find that the quickest and easiest way to meet the new 
challenges that continually arise is to tap their own networks to find executives for their teams,^ or to 
rely on other members of the organization to help find them. My data show that, across all C-level and 
VP-level hires, founder-CEOs are the source of 49% of all hires — by far the biggest share. Tapping 
their own networks to find hires provides the benefits of comfort and access, but also business 
benefits: Founders who hire via their networks are able to build more coherent organizations and to 
attract employees who are more receptive to the organization's control systems, freeing up the 
founders to focus on non-HR issues.^ When founder-CEO Genevieve Thiers began to hire people for 
her startup, an online babysitting site called Sittercity, she used her personal network because "hiring 
and managing people is a new thing for me. I definitely don't want to come out of the gate and have 
the wrong hire. So initially I decided to hire people I knew because it felt less risky." 

All three original founders of Pandora Radio, an online-music startup, preferred to staff their new 
company with friends and personal contacts, believing such employees would take more 
responsibility, make more sacrifices, and be more proactive for the company. Founder Tim 
Westergren explained his thinking about hiring friends: "I'm a huge believer in hiring people you 
know. Friends will be the ones who go to the mat for you, do it out of loyalty, and would be in the 



boat with us, as opposed to just being enployees. We were able to engender that sense of ownership 
in a very strong way." Research seems to bear him out. When hiring nonfounding executives, high- 
potential startups that relied relatively heavily on the founders' personal networks received 
valuations that were 37% higher than those received by startups that barely tapped the founders' 
networks.^ Playing such a central role in hiring executives can give the founder-CEO a closer initial 
relationship with his or her direct reports, though at the cost of having to spend more of his or her 
own time recruiting those people than if an HR function or third-party recruiter were handling the 
task. 

CEOs tend to feel more comfortable with hires they already know, especially compared to the 
hires who come from the next most common source — the startup's investors and board members. 
During the earliest stages of growth, investors are the source of few hires, but by the second round of 
financing, the founder-CEO is finding fewer members of the team and investors account for 19% of 
executive hires. Hires found through investors often have no tie to the CEO but a strong tie to the 
investor. Such hires may, in fact, be more aligned with the investor than with the CEO for whom they 
are supposed to be working, introducing potential agency issues and loyalty challenges that the 
founder- CEO has to manage. 

This is particularly true for CFOs. Investors are a source of 26%) of all CFO hires, a far higher 
percentage than for any other non-CEO position. If a CEO's network does not include solid CFO 
candidates, the investors may be playing a valuable role by finding CFOs for their startups. Some 
founders, however, take a darker view of the preponderance of investor-chosen CFOs; one referred 
to them as "the investors' eyes and ears, their spies on the team" Another founder observed that, 
while the founder is still CEO, the CFO position is one of "the main levers for investors" who focus 
on "picking a CFO who can keep tabs on how their capital is being used and on the startup's 
performance." In contrast, investors apparently focus much less on building ties to other potential 
executive hires, such as CTOs (for which investors serve as the source of only 13%o of hires), COOs 
(16%), and the layer of VPs below them (14%). 

Of executive hires, 15% are found through a non-CEO member of the executive team; these hires 
do not have a strong tie to the founder-CEO but do have a strong tie to one of the CEO's subordinates 
or to another member of the founding team When James Milmo, cofounder and chairman of Lynx 
Solutions, first started building the Lynx team, he and the founder-CEO realized that they were too 
busy to handle all of the hiring, "so we set out to hire someone to do the hiring. I got in touch with a 
friend from [college] whose judgment in people I really trusted. He had never done HR stuff before, 
but I asked him to be our head of hiring. . . . That was our most important hiring decision because he 
could hire a lot more people than we could. Plus, he did a great job of extending our hiring 
philosophy." 

The founder-CEO's centrality in hiring diminishes as the startup grows. Figure 8.1 shows that, 
before startups have raised any outside financing, their founder-CEOs are the sources of two-thirds of 
their C-level and VP-level executive hires. But as the startup matures, this percentage drops, possibly 
because the founder-CEO is running out of good candidates he or she already knows. As Genevieve 
Thiers of Sittercity observed, "I realized quickly that, to get the best people possible, I would have to 
go outside my network." Also, at this point, new players — investors and other outside directors — 
start bringing in hires with whom the founder-CEO does not have a relationship. By the time the 
second round has been raised, the founder-CEO accounts for fewer than half of all hires; this 
percentage continues to drop in subsequent rounds but remains higher than that of any other source 



through the fourth round of financing. 



Casting a Wide Net versus Getting Cultural Fit 



After the Lynx founders had tapped their personal networks, they began casting a wider net. For 
instance, they placed want ads in newspapers: "Our job postings stood out. They seemed like they 
were not written by HR people. [One of our early hires] was a Columbia B-school intern who we got 
to drop out of school and join us fiill-time. He said the reason he applied to us was he saw our ad that 
said you could 'work in shorts.' It caught his eye." More broadly, hires can come from a variety of 
weaker ties, including executive- search firms, want ads, cold submission of resumes, and other 
sources beyond the networks of the founders and other participants in the startup. As shown in Figure 
8.1, these weaker- tie sources of hires are important throughout all stages of startup evolution, 
accounting for between 14% and 29% of hires during each stage and averaging 20% across all stages. 

For job seekers of all types, weak ties can be a path to higher-status jobs.^ For founder-CEOs, though, 
relying on such weak ties can make it hard to judge an individual hire's cultural fit with the rest of the 
organization and, overall, hard to keep turnover low.^ 




30% 



20% 



10% 




<J/0 


Pre- 
funding 


After A- 
round 


After B- 
round 


After C- 
round 


After D- 
round 


^^CEO was source of hire 


66% 


58% 


49% 


47% 


40% 


Non-CEO executive was 
source of hire 


6% 


10% 


15% 


20% 


19% 


^if-lnvestof was source of hire 


6% 


11% 


19% 


19% 


13% 


^•Other source 


22% 


22% 


18% 


14% 


29% 



Figure 8. 1. Sources of Executive Hires in Startups Led by Founder-CEOs 



As the team grows beyond the founders, its dynamics change dramatically. The tight-knit founding 
team has to integrate executives and employees who may have very different motivations (they may, 
for example, see themselves as agents rather than stewards*) and skills (they may, for example, be 
much more specialized), but without whose contributions the startup cannot take the next step in 



growth. At the same time, new hires also have to be prepared for the challenges of integrating into a 
tight-knit founding team As do other groups of people who work closely on challenging tasks, 
founding teams often develop informal routines, processes, and shortcuts that outsiders find hard to 
understand. Dick Costolo, founder-CEO of FeedBurner, observed about his cofounders, "Early 
employees can be shocked by the brutal openness and the less-than-courteous exchanges between 
people who've worked together for 11 years, and you have to manage the introduction of these people 
with clear messages such as, 'If we start annoying you with verbal shorthand, just remind us that 
we're not the only people here anymore and to please elaborate.' " 

Evaluating a hire's fit with the startup can be much harder than evaluating his or her skills. At 
FeedBurner, Dick learned this lesson the hard way, but later saw parallels to his prior experience in 
improvisational comedy teams. Early in the startup, he recalls, "We had hired people who looked 
great on paper: the guy who had all the right credentials, great grades, knew all the programming 
languages. But during an interview, I often got the feeling, 'Not really one of us.' At [my prior startup] 
Spyonit, we would have hired him anyway, but seeing the problems it caused reminded me of the 
problems I had seen with the team of 'improv' all-stars, where the fit between personalities was so 
much more important than just finding people who [were very good comedians in their own rights]." 

The Playing- with-Flre Gap: The Risks of Hiring Friends and Family 

Given the comfort with people one already knows and the greater likelihood that they will fit in 
together, founder-CEOs find it very tempting to fill their teams with such hires. But this comfort 
comes at a cost, for such a team faces the same "playing-with-fire" risk that we saw founding teams 
facing in Chapter 4: Teams whose members have prior personal relationships may be less likely to 
discuss sensitive issues and may also face major damage to those relationships if things go sour in the 
startup. Recall that Evan Williams's team at Blogger included fi-iends and former colleagues; when 
Blogger ran out of money, Evan not only lost employees, but also lost his social circle. 

Tim Westergren enthusiastically hired friends at Pandora, but also experienced the downside: 
"You're mixing friendship with business. [At Pandora], we had to make a lot of thorny choices that 
negatively impacted our employees. The friendship had to come second because we had a higher 
responsibility to the collective. [But even though we made that choice,] we [still] suffered for that 
person. One of the quintessential questions for a manager is: If you are going to have layoffs, when do 
you tell people? When they are a friend you can see the absolute conflict, you want to give them fair 
warning — you want to give them all this time. But that's not the right answer for the company. So 
what do you do? It's a 'Sophie's Choice' — there is no winning answer for you as a person. You 
either do something that is against [your friend's] interest or against the interest of the broader 
conpany." 

Hiring people who have strong ties to other cofounders can introduce similar problems in dealing 
with underperformance. One founder related, "I made the mistake of trusting my founder/friend to hire 
the right people for operations. He oversold his family members' abilities and I ended up in a 
conpany with people who are virtually inpossible to fire and who get paid far more than their skill 
set is worth. They also don't want anyone rocking the boat for them." Another founder said, "After 
founding a software conpany with a friend who went and hired his family, my firsthand experience is 
that the dynamics of friendships — and worse, family — in a company can really create some miserable 
situations for people who are not part of the family power clique." 



ROLES 



Founders tend to start out with a flat, top-heavy organizational structure (described in Chapter 5), 
taking C-level titles for themselves and hiring few, if any, people below them to whom they can 
delegate tasks. However, as the startup matures, it gains more resources and becomes more 
formalized, enabling the founders to hire people below them to whom they can delegate tasks and thus 
make better use of their own time. They can gain efficiency, but must also face two major challenges 
described below: creating new C-level and VP-level positions that did not exist before and 
"upgrading" existing positions by replacing founders and early hires with hires more suited to the 
new demands of those positions. 

When to Create New Positions? 

Figure 8.2 shows that, as startups mature, executive teams go through two major changes. The first is 
a steady growth in the size of the team, from an average of about 3.5 executives to almost 6 by the 
time the startup raises its fifth round of financing. The second is the transition from top-heavy teams 
dominated by C-level executives (mainly founders, as described in Chapter 5) to teams in which VP- 
level executives finally outnumber the C-level executives after the fifth round of financing. But 
exactiy which of those positions will be created and when? 

At the C-level, the major new position is CFO. Only 4% of startups designate a founder as CFO 
(or head of finance) during their early stages. As startups mature and their financial challenges 
become more complex, they hire CFOs (or, less frequently, assign that position to someone already in 
the company); 70% of mature startups have a nonfounding CFO. Less common in young startups is the 
creation of the COO role (or, often interchangeably, the president role); before the first round of 
financing, 33% of startups have a COO/president (of which one-third are founders), but this 
percentage hasn't increased four rounds later. It may be that most young startups have little need for a 
COO/president beyond giving one of the founders a C-level titie or backing up an operationally 
deficient CEO. Bill Holodnak, head of executive- search firm J. Robert Scott, says, "If a startup has a 
COO, it's a red flag: Either the COO doesn't belong or the CEO doesn't." Bill's statement paints all 
startups with a broad brush, but captures the need for startups to evaluate their teams carefiilly before 
adding such positions. 




Figure 8.2. Changes in Executive Teams as Startups Mature 



As suggested by Figure 8.2, many more new positions are created at the next level down, where the 
vice president title is typically used. The first VP-level position created is often the VP-sales; even 
before raising the first round, 37% of startups have a VP-sales (4% have a founder in that position, 
33% have hired a nonfounder to fill it). The next most common early-stage VP position is VP- 
marketing (found in 22% of pre-financing startups), followed by VP- engineering (20%) of startups), 
VP-business development (19%o), and VP-human resources (19%)). As startups mature, gain 
resources, and formalize, they steadily add these specialized roles. After the fourth round of 
financing, 51% of startups have a VP-sales, 44%) have a VP- engineering, 37%o have a VP-marketing, 
and 36%o have a VP-business development. (Other startups may be using below- VP people to fill 
these roles. In addition, in three-quarters of the fastest- growing companies in the Inc. 500, a founder 
still served as the company's chief or only salesperson, even though the founder's title often did not 
explicitly acknowledge that role.)^ 

The one exception is the VP-human resources; even after the fourth round of financing, only 20% of 
startups have one. The team behind Spyonit and FeedBurner can help us understand why. At Spyonit, 
the team hired a VP of human resources early on, but in their subsequent startup, FeedBurner, they 
chose not to delegate that function. Founder-CEO Dick Costolo explained the evolution of the team's 
thinking on this issue: "[At Spyonit], none of us wanted to run hiring, so we brought in a person to run 
HR when we were only six people. A year later, we're looking at the code, asking why a certain 
component has so many bugs, and only then did we realize that we had hired several people who 
weren't cutting it! . . . The HR guy was telling us, 'This is a good person.' ... A year in, we realized 
we should not have hired the guy. . . . Hiring decisions became the HR guy's decisions, not the 
team's." Dick said that at the team's next startup, FeedBurner, "We asked, 'What part of the General 
& Administrative [functions] do we want to offload first?' and this time we decided to offload 



finance instead of HR. . . . We wanted to do all the interviewing ourselves this time and it made it 
easier to have someone take responsibility for the finance stuff." Other startups combine the HR 
responsibilities with other tasks rather than having a person dedicated to HR, or use junior people to 
perform the administrative HR tasks. 

As startups mature, they also begin hiring aggressively below the executive level. One metric we 
can use to examine this evolution is structural leverage — the number of nonexecutive employees per 
executive. As shown in Figure 8.3, startups begin life with very low structural leverage, about 2.5 
employees per executive, which steadily increases as the startup matures. By the time the startup 
raises its fifth round of financing, there is an average of 9.0 employees per executive, a situation with 
very different managerial challenges than the earlier low leverage. If an organization's tasks are such 
as can be delegated, and if the senior executives manage their employees effectively, increasing 
structural leverage can increase the firm's performance as junior enployees provide support for 
senior executives while learning best practices from them 



S 
I 



E 

> > 



8 

% 

e 

S 



10.0 
9.0 
8.0 
7.0 
6.0 
5.0 
4.0 
3.0 
2.0 
1.0 
0.0 







4.3 
































After A-round After B- round After C-round After D-round After E-round 

Figure 8.3. Changes in Structural Leverage as the Startup Grows 



The hiring process itself can force the organization to become more formalized. Creating job 
postings is often the first time a startup has to concretely define job requirements and think about the 
differences between various positions or roles, making such postings the startup's first real job 
descriptions. Each new hire also puts added pressure on the organization to be more explicit about 
processes and decision criteria, especially when the new hires have prior experience with formalized 
organizations and yearn for the additional clarity that can come from formalization. Dick Costolo 
relates how, at FeedBurner, "We brought in a person on the West Coast to run partnerships when 
Steve and I were both spending time on partners, and this person ended up having some issues with, 
'Who am I reporting to, you or Steve? I'm getting different answers from each.' As we start to grow, 
we can't have a loose oligarchy; we need people to be in charge of different things. . . . Pieces of the 
organization that we wanted to be very fluid, other people wanted to force them into having more 
structure." Alternatives to ordinary hierarchy, such as matrix structures or dual-report arrangements, 
can work in startups, but they introduce an even greater need for explicit discussion about roles and 
responsibilities. 



When Should I Upgrade Existing Positions? 

As startups grow, the demands on each position also tend to grow. For instance, the CTO whose job 
at first was to perform initial system design and development eventually becomes responsible for 
hiring and leading a technical team, which requires very different abilities. Likewise, the VP-sales, 
who began simply as the first salesperson, later finds himself or herself called upon to hire and lead a 
sales team and design a sales process and sales-compensation system, responsibilities that are very 
different fi"om selling. Some founders and early employees are able to grow and develop as the 
demands of their positions accelerate and as they need to change from "player" to "coach." But in a 
fast-growing startup, the demands will grow more quickly than most people can learn and the startup 
must decide whether to leave a crucial task in the hands of a founder or loyal early employee who 
can't really handle it or to replace that person with someone who can handle it. 

By any normal business standard, upgrading the quality of the person in the position makes more 
sense. Indeed, many startups do upgrade positions as they grow; in all C-level and VP-level 
positions, the percentage of founders is lower after the third round of financing than it was after the 
first round of financing.* 

When some founder-CEOs sense that one of their employees is underperforming, they immediately 
act to remove the person. One said, "Knowing/or sure that someone has to go is hard, but I have 
learned that if I start thinking someone needs to go, they need to go. It is always the right call to 
upgrade when you realize someone can't or isn't succeeding." 

However, the decision to replace an underperformer is not always as obvious as it might seem. In a 
startup, the shock to a small, tight-knit team could be considerable. In many cases, even when the 
damage an underperformer is doing to the company's value is clear, the founder-CEO or other 
founders simply can't bring themselves to throw one of their own overboard. Particularly when the 
division of labor is not very clear, the lack of individual accountability can be a trap. It can be too 
hard for the founder-CEO to justify an underperformer 's dismissal to others on the team, or it can be 
too easy for the CEO to convince himself or herself that the underperformer isn't really doing that 
badly. Founders may not even be able to bring themselves to discuss the problem (and possibly find 
some other solution) — a perfect example of the playing-with-fire risk associated with teams whose 
members have prior personal relationships. But keeping the underperformer may also have real 
advantages, allowing the startup to retain valuable knowledge, connections, and relationships and 
showing other employees that loyalty is reciprocated. Upgrading can leave other team members 
disillusioned with the costs of growth; startups are often populated by people who value the mission 
and the camaraderie more than maximum profit. In addition, the more idiosyncratic the fiinctional 
background of the first holder of a newly created position, the harder it is to find a replacement who 
will be able to stay in that position for a long time,^^ increasing the risk to the organization. 

Such situations enphasize the dangers introduced when a founding team adopts early C-level tities. 
Title inflation may be good for attracting potential cofounders or early hires to the startup, but it can 
stand in the way of upgrading the team when that becomes necessary. According to Barry Nails, 
founder-CEO of Masergy, "Early on, there's the CEO, and then you have to decide who else is hired 
at what titles. Do we hire him as a manager or as a VP reporting to the CEO? Early in the company's 
life, you don't have the ability to give them more salary, but titles are cheap to give out so you use that 
to make the package more attractive. They're currently a 'manager' in their company, so if they come 
in as a 'director,' they'll feel as if they've gotten a promotion. But, then they can't scale with the 
startup and that inflated title causes big problems." The person who enjoys having a senior title 



naturally resists being pushed a level or more downward by a new hire. This certainly caused 
problems for Nails in his own high-growth company: "Our VP of sales started off managing two 
people the first year. Responsibilities quickly expanded the following year to managing people across 
the U.S. and, the year after that, to managing people in London and international sales. After three 
years, it was evident I needed to hire a 'true' VP of sales, so I hired a 'senior VP' above him. I 
realized that, in every place I could, I had to set expectations and explain to someone who was being 
hired as a direct report to me that, 'At some point in the ftiture, I'll be hiring a boss for you. You'll 
have the ability to conpete for the position, but someone else fi^om outside the company could take the 
job.' Saying that doesn't quite work — anyone aggressive at all wants to believe they can work at a 
higher level. In fact, that aggressiveness is why we want them in the first place, but they usually don't 
scale." Clear communication and regular expectation setting are tough to accomplish, but a startup 
needs to have them. Even when it does, though, management challenges usually remain. 

Whom to Hire When? 

When deciding whom to hire into each of these new or upgraded positions, founders face at least two 
major trade-offs: hiring generalists versus specialists and hiring inexperienced versus experienced 
people. What's more, the trade-offs shift as the startup gains resources and becomes more formalized. 
Making the wrong decision at a particular stage of startup evolution can cause major problems, either 
immediately or down the road. 

Generalists versus Specialists: Option Value versus Depth 

Each time a startup hires a new employee, it faces a choice between hiring a specialist, who can 
usually be counted on to do a specific task well, or a flexible generalist or jack-of- all-trades, who 
cannot do any particular task as well as a specialist could but can move across multiple tasks more 
effectively than a specialist could. 

Which choice is best depends very much on how formalized the ftinction in question has become, 
which is dependent both on the startup's stage of development and on that specific ftinction. When 
startups are founded, their founders often have a solid initial idea of what product or service they 
want to develop and sell. However, they face many uncertainties about their strategy for doing so, the 
business model they will use to make money, and even what the oft-changing idea will look like 
several months from now. They resist formalizing processes or creating specialized roles, allowing 
each person's tasks to change daily to match the fluidity of the startup's strategy, business model, and 
product-development efforts. At this early stage, flexibility is a priority. 

Early-stage startups therefore tend to seek generalists who can pitch in wherever needed. Even if 
a hire is assigned to a specific ftinction, there is high "option value" in being able to move that person 
to another ftinction as the work dictates. Founder-CEO Dick Costolo said that, of FeedBurner's first 
20 employees, "In the end, five of our original twenty moved into radically different positions at 
some point." This flexibility is particularly important for ftinctions that are still fluid or still face 
important uncertainties. For instance, if developing the product requires the team to resolve important 
technical or scientific uncertainties, then the technical or scientific ftmctions will not be as stable and 
well defined as they are in startups for which the technology or science are well defined but there is 
much uncertainty about the market and its need. 



During this fluid stage — and especially for functions that are still evolving — hiring specialists who 
excel at specific tasks but cannot, or would not want to, contribute to other tasks can be a big mistake. 
Hiring the world's greatest cellist could come back to haunt you if you're not sure yet whether you are 
an orchestra or a marching band and, if you turn out to be a marching band, the cellist or the band will 
be in for a tough time. Dick Costolo learned this lesson the hard way in his first startup, DKA, when 
the team hired an experienced and aggressive VP of sales, who immediately requested a 
demonstration version of the product to help the sales effort. The problem was that the product had 
not yet solidified enough for a true demo. Dick explained, "The user-interface people . . . put stuff in 
the demo that we couldn't implement later. So the customers saw things we could only do in a demo 
environment and it led to inevitable customer disappointment when they didn't get it in the product 
itself It was a double negative; it also distracted our software people from getting actual product 
done for us to sell. It was a major time sink and opportunity cost." The team tried to slow the sales 
effort and redeploy that VP elsewhere, Dick recalled, but "we had hired a guy who was specifically a 
VP-sales and couldn't be anything else." 

After that costly experience, Dick placed a premium on hiring flexible people in the early days of 
his subsequent startups: "We might be hiring for a specific role, but we . . . look for someone who 
could go do other things if the company heads in another direction in the next 5-6 months." 

However, as major uncertainties are resolved and the startup begins locking down its technologies 
and strategies, flexibility becomes less valuable. The option value of being able to move employees 
from area to area is far lower and the value of having a specialist excel at a task is much higher. Over 
time, this accentuates what was earlier a less obvious cost of hiring generalists instead of specialists 
— the lower quality of the generalists' work. Drawing an analogy to athletes, Dick Costolo said, 
"With our early finance hire, it was like he had played wide receiver in college and we thought he 
could play defensive back in the pros so we drafted him. Then, we didn't use him as a defensive back 
and he ended up being only the third-best wide receiver. Oh well!" 

As the organization grows, it usually tries to improve its efficiency by starting to formalize 
processes and structures. Dick recalled, "Our first CFO always wanted to have org charts for the 
board because they add a lot of structure to the board [presentation], but there was no org chart until 
the end of 2004. I was perfectly happy month to month moving people around when we didn't find 
people we wanted to hire." FeedBurner created its first formal organizational chart when it passed 20 
employees and began to hire for specific and permanent positions. Mirroring a common pattern, 
FeedBurner 's technical functions were the first to become more specialized: "The business needs 
more of an emphasis on specialists now. In engineering, we need great Java architects who can jump 
into someone else's code and understand it and who specialize in writing great Java code. We've 
started to say, 'We need people who eat, drink, and sleep Java code.' I don't care anymore if they can 
do C++ code or if they can manage people. . . . Our job requirements have gotten very specific." 
During this transition, candidates' depth and expertise gain ascendance over breadth and fiexibility, 
but not necessarily at the same rate for all functions. At FeedBurner, "The other functions are 
different; they're still like we were in the beginning. . . . With finance, operations, and business 
development, the culture is still about being able to juggle lots of things at once, being able to 
multitask." 

It is often hard for founders to judge when to make this transition, to understand how it might differ 
by function, and to manage the disruption it might cause. James Milmo of Lynx Solutions refiected on 
how this transition affected Lynx: "What you don't realize until afterwards is that the early phase 



requires a different type of person. ... At some point, a business begins to gel, [then you] want 
people who are less creative, less curious, and are fine with sticking with a job for a while. . . . [For 
us,] migrating to Phase Two quickly was tough." Likewise, the SPEC study of emerging conpanies 
found that transitioning from more fluid blueprints to a bureaucratic blueprint caused turnover to 
jump, heightening the challenge for the organization.^^ 

Small-Company versus Big-Company Backgrounds: "He Can 't Actually Build the Crank" 

A related trade-off is the one between hiring people with experience at a big company, where they 
often become more specialized and get used to working in formalized organizations, versus people 
who have worked for startups or small companies and might have developed the broader skills of a 
generalist. Many of the pros and cons of this decision mirror those of potential founders deciding 
whether or not to build a career in a big company before becoming founders, as described in detail in 
Chapter 2. On the one hand, people working in big companies have an opportunity to learn processes 
and systems in a company that has succeeded well enough to grow large. On the other hand, many of 
the habits people develop in large companies can be counterproductive in a startup. As Frank 
Addante explained about one of his hiring mistakes, "We hired [a VP of sales] through a recruiter 
who was focused specifically on VPs of sales. He had worked at Oracle and at IBM. He looked great 
on paper. We brought him in and he talked a good game. Then he just sat there for three months. He 
didn't do anything. He didn't hire anyone. He didn't come up with a plan, he didn't create a strategy. 
He wasn't comfortable creating something from nothing. It's like if you have a crank, he can crank, 
but he can't actually build the crank. Building something from nothing requires a different skill set." 

Barry Nails learned about the related divide between "doer" and "manager" with an early hire at 
Masergy: "My first mis-hire was someone who . . . had done sales and sales management at a mid- 
sized-to-larger company, where he had a great team behind him. When you hire senior people in 
larger companies, they're successfiil because they can manage a team They won't necessarily be 
effective on an individual level. I realized that in an early-stage company, there's no such thing as a 
manager. Everyone is a contributor, including the CEO! People in mid- management and upper- 
management in large companies, when they have to go back five or six years to find a project they 
were proud of contributing to, that's a red fiag." 

Even if the hire is actively working on the startup's tasks, it may be on the wrong tasks or in ways 
that are counterproductive. Early in the life of Zipcar, founder Robin Chase felt she needed a more 
experienced executive on her team in order to raise capital. But hiring a seasoned manager caused 
more problems than it solved. She said, "Our mistake: hiring a big-company guy for a startup. He 
spent a lot of money on lunches and parking, created huge lists and detailed tasks and procedures that 
were 25% out of date by the time they hit my desk and 50% out of date by the following day. He was 
used to working at a much later- stage company where the goal was to put procedures in place and 
follow them strictly." 

When Frank Addante started Zondigo, a wireless advertising conpany, after his successfiil exit 
from L90, he wanted to assemble a "dream team" of people experienced with marketing and 
technology. He hired people who had held senior executive positions in established companies such 
as UPS, Visa, Coca-Cola, Columbia Pictures, and Intel. But his dream team failed to live up to 
expectations. As Frank explained, "They spent so much time thinking and pontificating and 
strategizing and analyzing, they could never make a decision together on how to move forward. They 



just kept debating. . . . The results were that nothing really happened." 

Experienced versus Inexperienced Hires 

The choice of generalist versus specialist is largely driven by the startup's emphasis on flexibility 
versus formalization. A related but distinct decision is how experienced each hire should be; this 
decision is often driven by how much experience the startup can afford. As shown in Figure 8.4, 
startups tend to hire relatively inexperienced employees before they have raised capital; that is, while 
they are still resource-poor. As they mature and gain additional resources, they tend to increase the 
seniority of their hires, in terms of both years of work experience and prior seniority in the given role. 
The benefits of hiring experienced people include the following: 

• Skills, contacts, and credibility — Experienced hires are more likely to bring human and 
social capital to the startup and to provide it with reputational advantages. 

• Hiring leverage — By hiring experienced people, then delegating to them the task of hiring 
their own employees, founders can both leverage their own time and foster cohesion within 
each department in the startup. For example, one of the main reasons Frank Addante hired 
experienced people at Zondigo was because he "looked forward to relying on them to hire the 
people who would be working under them" and thus fi^ee Frank to pursue ftind-raising. 

• Stability: Mr. Right versus Mr. Right Now — Experienced hires lower the probability of 
having to upgrade the team in the long run; their relevant experience usually helps them scale 
more smoothly with the growing organization. If a startup can attract a Mr. Right, who will be 
able to excel in the position through multiple stages of growth, rather than a Mr. Right Now, 
who can do the job now but will be out of his or her depth in the next stage of growth, that 
will save the startup a lot of trouble and disruption. Barry Nails of Masergy was surprised by 
this aspect of startup growth: "High-growth companies change so much every quarter! That 
amount of change isn't normal and people can't be trained for it unless they've been in another 
[startup or small company]." 



21.0 



20.0 




65.0% 



60.0% 



55.0% 



50.0% 



45.0% 



40.0% 



- 35.0% 



•Years of prior experience, 

C-level hires 



•% of C-level hires with prior 
senior experience in role 



Figure 8.4. Seniority of Hires, by Maturity of Startup (number of rounds of financing raised) 



At the same time, there are also downsides to hiring experienced people. They include these: 

• Bigger paychecks — ^At their current jobs, experienced candidates are probably earning more 
than junior candidates, making it more expensive to lure them to the startup and increasing the 
startup's "burn rate." Letting payroll get out of hand can cause not only dissatisfaction and 
turnover but also lawsuits. When Pandora Radio couldn't afford to pay its employees and had 
to defer salaries, it not only suffered a "pressure cooker" work environment, according to 
founder Tim Westergren, but also was sued for violating California employment laws. 

• Cultural control — With bottom-up hiring, founders will be more in control of the culture they 
want to create. People lacking years of experience at established companies do not have such 
strong expectations of how a company should run and so are more open to the founder's 
blueprint. James Milmo wanted his technology startup. Lynx, to be a home for creative 
"Renaissance" people; that is, people with high intelligence, creativity, and wide-ranging 
interests, but not necessarily with formal training in computer programming. James explained, 
"We hired a lot of Renaissance people who had backgrounds that were unrelated to what they 
would be doing for us. . . . Our culture was such that we got along with each other, spent a lot 
of time outside work together. We were a community. . . . People were more excited to be at 
Lynx because they loved who they worked with." 

A startup hires inexperienced people hoping they will prove to be "rising stars" who can master 
new skills and grow into their roles. In contrast, experienced hires are expected to be "rock stars" 
who can contribute a lot of value right from the beginning. Genevieve Thiers, founder-CEO of 
Sittercity, explained how her ideas about hiring evolved throughout her startup: "[At first] I had 



brought in 'rising stars' — ^yoimg people who would be able to grow with the company. But they were 
simply too emotional about the business swings of the company. It took too much of my time and 
energy to manage them. I felt as though I needed to find more senior 'rock stars' to help me take the 
conpany to the next level." She decided to focus more intensely on hiring a new executive team to 
replace those she had hoped would be rising stars and began searching for rock stars — people who 
were established and successful in their fields. Over the next year, Sittercity was able to attract 10 
senior people by offering stock options and competitive salaries, by offering the chance to enter an 
up-and-coming industry at the ground level, and — being located in Chicago rather than in Silicon 
Valley or along Route 128 — by not having to compete with a lot of other startups for local talent. 

While Genevieve gave up on rising stars in favor of rock stars, other founders were able to "hit the 
hiring jackpof by finding that rare commodity, the young, relatively inexperienced (and cheap) 
person who can learn and grow quickly to become a true asset to the startup. As Lew Cirne was 
ramping up Wily Technology, one of his first hires was Mark Sachleben, a recently minted Stanford 
MBA and a friend of Wily's outside board member. Lew was a little worried about bringing in an 
MBA at such an early stage, but felt that Mark's youth, potential, and working style might be a good 
fit with his young company and decided to make him CFO. "Mark was my gold-plated Stanford 
MBA," recalls Cirne, "though I wasn't able to pay him much gold. I really liked that he had 
incredible integrity and a self-effacing nature. . . . Rather than buying already-built desks, I remember 
my Stanford- MBA CFO hand-assembling the desks we had bought, rather than paying $50 for Office 
Depot to assemble them." As Wily grew, Mark proved to be a quick learner and a key member of the 
management team 

REWARDS 

Cash-poor startups usually cannot afford to pay salaries and bonuses comparable to those paid by 
large companies. To compensate, startups have to use other financial (and non-financial) 
inducements, the most prominent of which is equity. Although the compensation packages of founders 
are often a mixture of incentives appropriate to both agents and principals, hires are usually classic 
agents who require different incentives. When designing compensation packages, startups can choose 
from contingent alternatives (i.e., performance-based bonuses) and noncontingent alternatives 
(salaries) and can tie financial rewards to individual performance (i.e., a bonus tied to the specific 
employee's performance^^) or to the startup's collective performance (most prominently via equity 
stakes in the startup). The paragraphs below describe the drivers of each form of nonfounder 
conpensation and then delve into some of the trade-offs between the different forms. 

Cash Compensation 

Nonfounder cash compensation can include both salary, paid regardless of individual and startup 
performance, and bonus, which is performance-dependent. Below, I describe the core salary and 
bonus patterns in the technology industry, then compare them with patterns in the life sciences 
industry. 



The Ups and Downs of Salaries 



Over the past decade, the technology and life sciences industries have experienced two full market 
cycles, from the heights of the economic boom in 1999-2000 to the depths of the bust in 2001-2003, 
then back to heights by 2007 and the subsequent recession in 2008-2009. As shown in Figure 8.5, the 
changes in nonfounder-CEO salaries in technology startups have largely followed the overall 
economic cycle, though lagging it slightly. C-level salaries declined during the downturns of 2001- 
2002 and 2009 and rose during the period in between. 

Analyses of the salary data show that, across executives and startups, the major differences in 
salary are as follows: 

• Position — ^As a group, C-level executives receive higher salaries than do VP-level 
executives, but there are significant exceptions. Across the frill decade of compensation data 
for nonfounding executives, the two highest-paid positions were the CEO (average salary of 
$217,000) and the COO/president ($176,000). \n the next tier, CTOs, CFOs, and the VPs of 
engineering, business development, sales, and marketing all made between $147,000 and 
$155,000.* To the extent that salaries give us a window into which frmctions are most critical 
for mitigating the challenges or contingencies faced by an organization,^^ it is interesting to 
see the similarity of salaries across most of the executive team, a pattern that is found in both 
early-stage and later-stage startups. 

• Stage of startup — The more mature the startup (the more rounds of financing raised) and the 
larger the startup (the greater the number of employees and the greater the revenues), the 
higher the cash compensation. For instance, splitting my 2009 startups into those that had 
raised two or fewer rounds of financing and those that had raised three or more rounds, I 
found that the nonfounder C-level salaries were 12% to 17% higher in the later-stage startups. 
Splitting the startups into those with 40 or fewer employees and those with over 40, 
nonfounder C-level salaries were 17% to 19% higher in the larger startups. Splitting the 
startups into those with less than $5 million in revenues and those with $5 million or more in 
revenues, nonfounder C-level salaries were 10%) to 12%o higher in the higher-revenue 
startups. 

• Industry — Executives in life sciences startups tend to make significantly more cash 
compensation than their counterparts in technology startups. For instance, in 2009, average 
nonfounder-CEO salary for life sciences startups was $285,000, compared to $231,000 in 
technology startups, a gap of $54,000. For the other nonfounder C-level executives, the gap 
was $31,000 to $36,000, or 19%o to 22%), in favor of life sciences executives. Across the frill 
decade, non- founding life sciences CEOs received, on average, a salary 27%) higher than that 
of their technology counterparts.* 



$260 






2O0O 


2001 


2002 


2003 


2004 


2005 


2006 


2007 


2008 


2009 


♦ Wonfo under CEO Salary 


$200 


$212 


$195 


$217 


$216 


$211 


$220 


$227 


$237 


$231 


B Nonfounder COO Salary 


$175 


$175 


$171 


$168 


$178 


$173 


$177 


$175 


S183 


$186 


A Nonfo under CFO Salary 


$150 


$142 


$126 


$144 


$141 


$141 


$148 


$158 


$165 
S170 


$164 


M Nonfounder CTO Salary 


$150 


$155 


$135 


$144 


$147 


$149 


$155 


$158 


$162 



Figure 8.5. C-level Nonfounder Salaries, 2000-2009 



Interestingly, geography does not seem to affect nonfounder salaries. In the technology "hubs" of 
California and New England, non- founding C-level executives made only about 1% more than their 
counterparts in the rest of the country, a difference that is not statistically significant. 

High versus Low Bonus -Sensitivity 

For many executives, cash bonuses are an important component of cash compensation. Such bonuses 
are often tied to individual performance measures and to startup-level benchmarks, making them more 
sensitive than salary to actual performance. Across the fiill executive team, bonuses average 28% of 
total cash compensation. However, there are significant differences by position. Figure 8.6 shows 
bonus as a percentage of the executive's salary across the full decade of data for technology startups. 
Vice presidents of sales had by far the highest percentage of their cash compensation, 49%, * in the 
form of a bonus; for CEOs, the figure was 37%, and for all other executives, below 30%. 

After two turns as a founder-CEO, Dick Costolo concluded that the "ideal conpensation package" 
depended on the function: Salespeople were heavily motivated by performance-based condensation 
while software engineers primarily wanted to count on their monthly paycheck. As Dick explained, 
with salespeople, it's best to "make the compensation package based on how much more they'll make 
if they meet or beat quota and I, as CEO, have to be comfortable with their making four times as much 
as me, but with their having a low base and huge upside." 

Bonuses also differ by industry. Cash compensation tends to be more at risk in technology startups; 
bonuses for nonfounding technology CEOs are 37%) of their salaries, as opposed to 28%o for non- 
founding life sciences CEOs. Similar patterns obtained across the other positions. (As described 
below, the equity stakes of nonfounding life sciences executives were also lower than the equity 



stakes of their technology counterparts.) 




CEO coo ao CFO VP-Sales VP-Eng. VP-BD VP-Mktg. 

Figure 8.6. Bonus as Percentage of Salary for Nonfounding Executives in Technology Startups, across the Full Decade 



Gender Gaps: Male versus Female Compensation 

Studies of cash compensation in large companies consistently find a "gender gap"; that is, 20% to 
25% lower compensation for women.* Even though large, bureaucratic organizations standardize 
their HR practices to minimize the effects of personal preferences and discrimination on employment 
and compensation decisions, this significant gap remains, partly due to the fact that men and women 

receive different job tities and are sorted into different occupational lines. In pre-bureaucratic 
startups, lacking such rationalized HR practices, do we therefore see an even bigger gender gap, or 
are there counterbalancing effects? If there is a gender gap in startups, can it give us insight into the 
gender gap in large companies? 

My analyses show that, in startups, there is a much greater gap in the preponderance of women than 
in their compensation. Across the technology startups I analyzed, the percentage of women across the 
C-level and VP-level executives was only 10.7%); in life sciences, it was \1 .6%?^ Only 3.1%) of 
technology startups and 1.9% of life sciences startups had female CEOs, but the percentages rose for 
the next tier of C-level executives and were even higher at the VP level. Outside of human resources 
(more than three-quarters of the heads of HR were women), women were most likely to be the VP- 
marketing (in 19.1%) of technology startups and 41.9% of life sciences startups). 

The gender gap in compensation was just over $10,000, or 5.6%), much lower than the typical 20%o 
to 25%) in large public companies. There were exceptions, however, which hint at the potential 
causes of the overall gap (and which suggest high-potential areas for research on the gender gap). The 
biggest distinction in gender gaps was geographic; the gap was $16,300 outside the entrepreneurial 
hubs of California and Massachusetts but too small to be statistically significant within them. 
Interestingly, as mentioned above, geography has no significant inpact on startup compensation in 



general. Why, then, does it have such a strong impact on the gender gap? One reason may be gender- 
related differences in the balance of (labor) supply and (job) demand in those two job markets. There 
may also be a voluntary element: Women who choose not to work in the hubs may be trading off 
condensation for benefits such as a more balanced lifestyle; if such benefits are more important to 
women than to men, they could cause a gender gap. 

Second, the startup's stage of development had a significant impact; there was a $12,300 gap in 
later-stage startups but no statistically significant gap in early-stage startups. This result has the most 
direct implications for our understanding of the gender gap in large companies, in particular for how 
it emerges and thus what the drivers are. Early in the life of a startup, there is no significant gap, 
suggesting that male and female executives are making similar trade-offs or that the startups operate 
as meritocracies. However, as startups grow, formalize their processes, or allow their employees to 
make more diverse trade-offs in compensation or lifestyle, a gap is introduced. Small at first, it 
widens as startups grow into the kinds of large, public companies that have been studied in the past. 
One (male) serial entrepreneur observed, "Early-stage companies are meritocracies, whereas mature 
companies are bureaucracies that reward factors other than performance and may be biased against 
women. . . . Most founder/CEOs get replaced by 'seasoned executives,' many of whom come from 
larger organizations. And those executives bring not only their 'own team' but they also bring their 
existing preferences and biases."* 

The third big influence on the gender gap was the industry: Across all startups in the sample, there 
was a $13,300 gap in technology startups but no statistically significant gap in life sciences startups. 
One reason maybe the aforementioned difference in the preponderance of female executives; they are 
much more common in life sciences than in technology. 

If startups are indeed more meritocratic than large companies, making them relatively more 
attractive as places for women to work, we may see, in the fiiture, an increase in the preponderance of 

women in startups, which may help reduce the gender condensation gap.^ To the extent that the gap 
also represents men and women making conscious and different trade-offs, the gap may even be a 
good thing, for people of each gender may be getting what is most important to them A female 
founder-CEO said, "I think women don't tend to leave because of salary whereas men do. A good 
work environment means a lot to us." However, to the extent that startup growth inherently nurtures 
the large, persistent, and involuntary gender gaps characteristic of large companies, it may be 
precisely during the early stages of startup evolution that we should be trying to find out how this 
happens and how to stop it. 

C-level versus VP-level Equity Stakes 

During the early days of a startup, equity stakes tend to be individually tailored to each new hire. But 
as a startup matures, this ad-hoc approach can cause trouble. Dick Costolo ran into this problem as 
FeedBurner grew: "Our CFO had been a venture capitalist and pointed out that we needed to [clean 
up all our ad-hoc equity arrangements with our employees] so all new hires would be in the same 
boat. . . . [EJveryone's interests weren't aligned." Dick and the CFO were able to bring consistency 
to the equity packages for subsequent hires, but the inconsistency among early hires ended up causing 
alignment problems when FeedBurner later received an acquisition offer. Employees with one equity 
structure wanted to take the offer while others did not, causing divergence within the team at a crucial 
juncture. 



Yet, even in the youngest of startups in my dataset, there are some consistencies. While the salaries 
of VP-level executives are often on par with those of some C-level executives (most notably, the 
CTO and CFO), equity stakes are much more strongly linked to organizational level. Nonfounding 
CEOs have significantly more equity than the rest of the C-level executives, who usually have more 
equity than VPs. Across the ftill decade of my dataset, nonfounding CEOs averaged equity stakes of 
6.0%, COOs 2.9%, CTOs 1.7%, and CFOs 1.3%. The four main VP-level positions all averaged 
1.0% to 1.3% of equity. 

Figure 8.7 graphs the equity and cash compensation differences across these positions, using as a 
baseline the VP-marketing, the lowest-paid and lowest-stake executive in the core executive team. 
Most members of the team earn within 5% of the VP-marketing's cash condensation; even the CEO 
makes only 1.7 times as much. The equity-stake differentials, however, are far greater; the CEO's 
stake is more than 6 times that of the VP-marketing. 

The stage of the startup also significantly affects equity stakes through dilution and uncertainty. 
First, each time a startup raises a round of financing, the equity stakes of existing equity holders 
decrease — they are "diluted" — because new shares are issued to the new investors. For instance, if a 
nonfounding CEO owns 10% of the startup and then the startup raises $5 million on a pre-money 
valuation of $10 million, the CEO's equity stake would be diluted to 6.7%). Second, during the early 
days of the startup, when uncertainty is high and hires are taking a big risk by joining, the startup has 
to offer larger equity stakes to attract them But as the risk is lowered, startups can offer lower equity 
stakes,* reinforcing the negative relationship between startup maturity and the size of hires' equity 
stakes. In my 2009 technology- startup data, the nonfounding CEOs of early- stage startups that had 
raised two or fewer rounds of financing averaged 7.1%) equity stakes; their counterparts in later-stage 
startups that had raised three or more rounds averaged 5.2%o. In life sciences, the gap was even 
larger: 2.5% instead of 1.9%. 




Industry also plays a small but significant role in equity stakes. Executives tend to have higher 
equity stakes in technology startups than in life sciences startups; in 2009, across all C-level and VP- 



level positions, technology executives each held an average of 0.5% more equity than did equivalent 
life sciences executives. All told, technology executives tend to have higher-powered incentives than 
do life sciences executives, with more of their condensation coming in at-risk forms of cash and 
equity: lower salaries, higher bonus percentages, and more equity. 

Locking on the Golden Handcuffs 

In Chapter 6, we examined the founding team's need to adopt dynamic equity arrangements, such as 
vesting schedules, in order to be able to adjust to future developments, such as a founder deciding to 
drop out or being forced by circumstances to curtail his or her commitment. The vesting terms used 
when assigning equity to hires give us a related window into equity arrangements within startups. 
Startups often use vesting schedules to handcuff a new executive; that is, to give him or her a strong 
financial motivation to stay with the startup for a certain number of years. (Vesting for nonfounder- 
CEOs is generally time-based, though about 10% of nonfounder-CEOs have performance-based 
vesting.) Ideally, each executive should be locked in for the horizon over which he or she is expected 
to add value to the startup. However, more than three-quarters of executive hires receive vesting 
terms of four years, as shown in Figure 8.8. The consistency of this arrangement across the various 
executive positions and throughout all parts of the market cycle, including boom times when startups 
are typically growing and reaching exits much more quickly and bust times when they may grow much 
more slowly, suggests that startups are following an institutionalized pattern rather than tailoring each 
vesting schedule to a specific executive and situation.* 

It is also instructive to compare the vesting schedules for founders and nonfounders; boards give 
founders an average of six months less vesting than they give hires. As shown in Figure 8.8, founder- 
CEOs are much less likely to have four years of vesting and more likely to have no vesting or only 
one year of it than nonfounder-CEOs are, even after investors have entered the picture. When faced 
with an investor demand that they vest their stock, founders often push to get "credit for time served," 
which may account for some of the six- month difference. However, the shorter vesting for founders 
may also be due in part to an underappreciated benefit of founder attachment to the startup. We saw in 
Chapter 6 that emotional attachment to the startup can handicap a founder's ability to negotiate an 
increase in pay, resulting in a founder discount. But the same attachment may actually benefit the 
founder when it comes to vesting; investors may feel less need to impose "golden handcuffs" on a 
founder they know would be most unwilling to leave. 




Cash-vers US-Equity Seesaw 

Different mixes of cash compensation and equity can create different incentives, encourage different 
levels of risk taking, and direct attention to corporate versus individual goals. Startups, like larger 
conpanies, often craft compensation packages that trade off cash compensation for equity stakes; if a 

package is high in salary, it will tend to be low in equity, or vice versa.^"^ Such packages are designed 
both to attract the right kinds of people to the startup and to provide them with particular incentives 
once they join. For instance, as a way to learn about a potential executive's risk preference. Bill 
Holodnak, head of the J. Robert Scott executive- search firm, shows those executives a spectrum of 
compensation packages, ranging from low salary/high equity to high salary/low equity, and asks the 
recruit where he or she would like to fall on that spectrum. 

Some startups choose to have all hires on one side of the seesaw or the other, while others allow 
each new hire to choose. Dick Costolo said that, at FeedBurner, "Each hire would be either an 
equity-heavy hire with below-market salary, or salary-heavy with below-market equity. When we 
would make people offers, I would ask, 'Do you want us to favor salary or equity?' For the ones who 
favored salary, I told them, 'Well, no complaints when it comes time to sell the company and your 
colleagues are going through the roof and you mortgaged your ftiture for more salary.' " Dick also 
learned that different groups sometimes have different equity- versus- salary priorities: "In Silicon 
Valley, the discussions are, 'Can you give me more options if I take less salary?' [In Chicago], we 
always have the opposite discussion with engineers." 

In analyses of my nonfounder-CEO compensation data, I found a strong and significant negative 
relationship between the CEO's salary and his or her equity stake, even after controlling for a wide 
array of startup and executive characteristics (such as the startup's maturity and industry, and the 
executive's position and experience). Splitting the sample of CEOs into the higher-paid half (average 
salary of $253,000) and the lower-paid half (average salary of $181,000), I found that the higher-paid 
had average equity stakes of 6.46% while the lower-paid had average equity stakes of 8.49%. In 
exchange for greater salaries, such CEOs gave up a significant amount of equity, or vice versa. The 



same relationship held for the other executive positions. 

Where employees come out on the cash-versus-equity seesaw can be affected by the founders' 
motivations as well as by the enployees' own. Founders who want to retain more equity for 
themselves tend to offer a smaller equity stake to potential hires, preferring either to balance that with 
some other carrot (higher salaries if they have the resources, or intangible benefits) or to take the 
chance of losing the hire and having to hire a weaker candidate. Such a founder is often able to retain 
more equity but may build a team that is weaker or less motivated to build equity value. The founder 
thus decreases his or her chances of ending up with a King outcome. On the flip side, founders who 
prefer to build a more valuable startup, even if their percentage of it is lower, will make the opposite 
decisions and increase their chances of ending up with a Rich outcome. 

CLOSING REMARKS 

Founders face the same dilemmas with hiring decisions that they face with cofounding decisions, but 
with the added complexity of aligning these decisions with the startup's stage of development, its 
level of formality (which may differ by fiinction, complicating the challenge even fiirther), and its 
available resources. Figure 8.9 summarizes the changes in hiring decisions that typically occur during 
each stage of startup evolution. 

It is crucial for founder-CEOs not only to diagnose and correct the problems that arise from hiring 
the wrong person, but also to anticipate when a decision that was good at the time may need to be 
rethought because the startup has arrived at a new stage of development. In the fast-paced world of 
startups, founders can't afford to become entrenched in a particular way of hiring but must expect 
their startup's hiring needs to change. For exanple, before hiring young enployees during the early 
days of the startup, the founders should already be planning for the day when they will have to either 
replace those young hires or else hire more experienced people to supervise them; before building a 
team of generalists, the founders should already be considering how they will adjust when the startup 
requires the quality that can be provided only by specialists; before crafting a compensation package 
for the earliest hires, the founders should already be anticipating how different stages will require 
different packages. As we have seen throughout this chapter, the data indicate that each stage of a 
startup's evolution is marked by important changes in its hiring needs. 



Startup's 
Stiigc i)f 



DeivUtpmeiil 


Rchtitmships 


Roles 


Rewards 


Startup 


• Personal networks 


• Gcneralists who 


• Low cash 




of core founder 


cover multiple 


compensation 




are tapped to find 


areas 


• High equity 




loyal candidates 


• "Fiat" structure 


compensation 




who fit with the 


that has many 


• Low gender 




culture of the 


C-level employees 






startup 


with few reports 


• Less vesting 


Transition 


• Impersonal 


• "Players" 


• Moderate 




searches (e.g.. 


transition to 


cash 




newspaper ads. 


"coaches," as 


compensation 




search firms) 


functional VPs 


• Lower equity 




• The networks 


are delegated the 


compensation 




(and weaker tics) 


responsibility- to 


♦ Vesting equity 




of investors and 


run and hire their 


stakes 




other participants 


own teams 






in the startup arc 


• Some early 






leveraged 


employees arc 
usually unable 
to adapt to the 
changing needs of 

fill* i*/^nmiiiv 




Mntiirc 


• Investor networks 


• The reporting 


• High cash 




are tapped 


structure is 


compensation 




• Executive search 


"pyramidal," 


• Employee 




firms arc hired 


with a few 
senit>r executives 
leveraged by many 
junior employees 
• "Profcssionar 
executives from 
large-company 
backgrounds 


stock options 
take the place 
of equity 
• The gender 
gap emerges 



Figure 8.9. Evolution of Hiring Decisions as the Startup Matures 

Another echo of the cofounder dilemmas can be found in the hiring dilemma of what to do when an 
underperforming or nonscaling employee is also a friend or family member. Founder-CEOs find it 
very tempting to fill their teams with such hires, given the natural comfort with people one already 
knows and the greater likelihood that they will all fit in together. But this comfort introduces longer- 
term risks, for such a team faces the same playing-with-fire risk that we saw founding teams facing in 
Chapter 4 and that we saw Tim Westergren facing earlier in this chapter. Founders who have prior 
personal relationships with their hires may be less likely to discuss sensitive issues and may also 
face major damage to those relationships if things go sour in the startup, both factors leading them to 
put off the issue until things are ready to blow up. 

Just as it was critical for founding teams to create appropriate linkages among their own 
relationships, roles, and rewards, it is critical for the founders to create such linkages with their 
hires. For example, a bonus-heavy compensation structure tied to individual performance may be 
productive with a salesperson but counterproductive with a programmer who is part of a 
development team; that is, rewards properly linked (or aligned) with one type of role may be poorly 
linked (misaligned) with another type of role. Creative linkages can help solve sticky hiring 
problems, such as the ripple effects of dismissing a hire who has turned out to be a poor choice. Dick 



Costolo of FeedBurner said, "The worst problem with a bad hire is that they get into the organization, 
they become friends with people, and then it's, 'Why are you firing my friend?' " The result is that 
many founders become hesitant to hire — or make sure to "hire slow" — for fear of making a mistake. 
But linking rewards and roles can help solve this problem; for example, a highly contingent 
condensation structure for salespeople can turn off those who aren't confident in their skills while 
encouraging weak hires to leave. As Dick observed, "A salesperson will leave if he's not doing a 
good job and therefore not getting paid what he wants. If he doesn't work out, it solves itself quickly. 
It's 'hire fast, fire fast' " Unfortunately, few first-time founders have such a nuanced understanding of 
how to balance hiring dilemmas; even for Dick, such understanding did not come until his fourth 
startup. 



* The three bases of recruitment were short-term skill set, long-term potential, and values/ attitudinal fit. The three bases of rewards 
were "love" or affiliation, "work" or a purposive drive to create something great with the company, and "money." The four means of 
coordinating and controlling work were "formal" oversight through job descriptions and HR procedures such as performance reviews 
and job descriptions, "direcf oversight through monitoring and influencing behavior, "informal" control through peer reinforcement and 
culture, and "professional" control through hiring from elite sources and assuming a degree of prior socialization in the employee. 

^ These five are (a) the "star" blueprint, with recruitment based on potential, rewards based on cutting-edge work, and control through 
elite socialization; (b) the "engineering" blueprint, with recruitment based on skills, rewards based on cutting-edge work, and control 
through peer enforcement; (c) the "commitment" blueprint, with recruitment based on fit, rewards based on love of the organization, and 
control through peer enforcement; (d) the "bureaucracy" blueprint, with recruitment based on immediate skill set, rewards based on 
work, and control through formal HR mechanisms; and (e) the "autocracy" blueprint, with recruitment based on immediate skill set, 
rewards based on financial remuneration, and control through direct oversight. 

* "Strong" and "weak" ties refer to the degree to which people are connected and have a prior interpersonal relationship. The strength 
of the tie depends on the amount of time people have been connected, the emotional intensity and intimacy of the relationship, and the 
degree of reciprocity. For more information, see Granovetter (1973). 

* See "The Founder Discount" in Chapter 6 for a brief description of agency theory versus stewardship theory. 

* In my dataset, it was lower by 4 to 17 percentage points, depending on the specific position and controlling for changes in the number 
of aU such positions. 

* A similar relationship holds for total cash compensation, which includes both salary and bonus. Over the decade, the average total 
cash compensation for a CEO was $296,000. For the COO/president, it was $228,000, and for the VP-sales, $230,000. At the next tier of 
nonfounder compensation, CTOs, CFOs, and the VPs of engineering, business development, and marketing all made between $180,000 
and $189,000. 

* There are some interesting divergences, but also parallels, between the compensation of CEOs in private versus public information 
technology and life sciences companies. Comparing cash compensation in private and public companies, the public-company CEOs make 
far more cash compensation than do their counterparts in private companies. According to the Wall Street Journal fRay Group 2009 
CEO Compensation Study, which analyzes compensation in public companies with more than $5 billion in revenues, CEOs of 
"technology" companies made $928,000 in salary and $1,084,000 in bonus and CEOs of "healthcare" companies made $1,200,000 in 
salary and $2,188,000 in bonus. These figures are much higher than the private-company compensation numbers shown in Figure 8.5 and 
described in the accompanying text. Slicing the pie another way and comparing cash compensation in information technology versus life 
sciences companies, the CEOs of public life sciences companies make significantly more than do the CEOs of public information 
technology companies, a relationship that also holds for private information technology versus life sciences companies. 

* Having such a large bonus component boosts the average cash compensation for non-founding vice presidents of sales into parity 
with nonfounding COOs, as detailed two footnotes above. 

* For details, see Goldin (2008). Gender gap studies have been criticized for failing to control for differences in human capital and job 
function. In contrast, in my analyses, I was able to control for a wide variety of differences in executive backgrounds, positions, 
industries, geography, and other factors. See Note 22 of this chapter for more details. 

* This observation implies that, if founders are still around, particularly in the CEO position, we might see less of a gender gap. To test 
this possibility, I added a flag for whether a founder was stiU working as a senior executive and reran my gender-gap regressions. The 
presence of founders did counterbalance the gender gap a bit, reducing it by about one-third, but the effect was not statistically 
significant. 



^ The CareerLeader data described in Chapter 2 — ^which suggest that women's motivations differ from men's, especially regarding 
the importance of financial motivations — lend some support to the suggestion that women like to strike different trade-offs than men do. 

* To provide sufficient equity with which to hire key personnel, many startups create option pools containing shares of equity that will 
be assigned to upcoming new hires. These are usually created or "refreshed" during each new round of financing, but 46% of pre- 
financing startups also create such pools in order to facilitate hiring. This percentage rises to 80% of startups when the first round of 
financing is raised, and to more than 90% from the second round onward. In principle, these pools should be sized to match the amount of 
hiring the startup expects to do before the pool is refreshed again; that is, until the next round of financing. The data do indeed support 
this relationship: The size of the pool is largest during the early stages of startup evolution, when the firm needs the largest equity stakes 
to attract hires and when the greatest number of important positions have yet to be filled. In the youngest startups in my dataset, the 
option pools averaged 20% of fully diluted equity; this percentage declined with each subsequent round of financing. By the third and 
fourth rounds, the option pool has dropped to 11% and 10%, respectively, reflecting the lower number of key hires remaining to be made 
and/or the smaller equity stakes required to attract them. 

* Another vesting term — one that can unlock the handcuffs of vesting — accelerates the person's vesting if the startup is acquired or 
there is some other change in control. There are various forms of this accelerated vesting on change in control term (e.g., different 
triggers that are required before the accelerated vesting occurs), but a common pattern is that the more valuable the employee to the 
startup, the less of a chance that the handcuffs will be unlocked, otherwise the terms of the acquisition will probably be less attractive to 
the acquirer (under the expectation that the most valuable employees will be able to leave). 



CHAPTER NINE 

INVESTOR DILEMMAS: ADDING VALUE, 
ADDING RISKS 



Just AS Evan Williams took widely differing approaches to his cofounding and hiring decisions 
at Blogger and Odeo, so he chose different kinds of investors for those two startups. In the early 
stages of Blogger, he and his cofounder, Meg Hourihan, financed the company with money they 
earned doing Web-development work for Hewlett-Packard. They hoped Blogger could avoid raising 
money and become "cash-flow positive" (when incoming cash is greater than outgoing cash) by 
increasing sales and software subscriptions, but Evan soon realized that they would need additional 
fimding to conplete the product development. With the dot-com boom heating up in 1999-2000, Evan 
could have approached venture capitalists, but he consciously decided not to do so because he 
wanted to avoid "losing a large percentage of our company." Instead, Blogger raised $500,000 on a 
pre-money valuation of $2 million fi^om small investors, including Meg's parents, in return for 20% 
of the company. "We were consciously not taking a lot of money. ... We were going to be frugal," 
explained Evan. "And the fact that we could still control 80% of the company was definitely 
appealing." When even that money ran out amid the bursting of the dot-com bubble, Evan resisted 
possible exits and instead financed the company by maxing out his own credit cards, eventually laying 
off all the employees and continuing to develop Blogger on his own. 

After Evan sold Blogger to Google in 2002, he helped found the podcasting startup Odeo. This 
time, he seeded the startup himself by using some of the Google stock he had sold, but he quickly 
recognized the huge potential for the technology and the risk that competitors such as Apple would 
beat Odeo to market. Because Evan's small-but-successftil exit at Blogger had made him a proven 
entrepreneur, he was able to raise $5 million from a top-tier VC firm which, in return, received 30% 
of the company and a 1.5 x liquidation preference.* The two founders would have two seats on the 
five-person board of directors. Although Evan had a $1 million offer from angel investors, the VCs 
offered him $5 million, and he decided to take the bigger check. 

Professional investors have played important early roles in financing and building some of the most 
successfiil startups, including such indusfry pioneers as Google, eBay, and Genentech. However, 
investors who are considering investing in a startup face many risks. The founder knows far more 
about his or her capabilities and motivations than the investors do and usually has a better 
understanding of the market potential. Investors can take actions, such as performing due diligence on 
the founder and the market, that will reduce their risks. But they must also structure the investment to 
reduce risk and to align the startup's interests with their own, so that "simply by following his or her 

self-interest," each party "advances the other's interests as well."^ Sometimes, the terms that are used 
in those investment structures (e.g., the founder vesting, liquidation preferences, and board 
representation detailed below) are necessary to make the investment possible, but, as we will see, 
also cause increased tensions between the founder and the investors. Such was the case with some of 



the terms to which Evan agreed when he accepted the VC offer. 
ENTER THE INVESTOR 

To survive and grow, startups need human capital, social capital, and financial capital. Core founders 
who lack any of these three often try to attract cofounders or hire employees who can provide what is 
missing. Sometimes that works. But for many founders, the startup's financial capital needs (e.g., for 
product development or for rapid growth) exceed the capital they can provide, and lead the founders 
to seek outside fiinding. When they do so, a brand-new player enters the startup — the investor — which 
can cause dramatic and sometimes unexpected internal changes. 

No other topic discussed in this book has received as much attention from academics as investor 
and financing issues have. While the academic research has not focused on intrateam financing issues, 
such as the equity splits covered in Chapter 6, "Reward Dilemmas," it has much to teach us about 
entrepreneurial financing that involves outside investors. It is beyond the scope of this chapter to 
review all of that research and to delve into the nitty-gritty of each financing term. Instead, I will 
build on the research in order to focus on the investor issues that result in founders' dilemmas: the 
early financing decisions that have long-term implications for growth and control. We will pay 
particular attention to building and managing a board of directors, often the conduit through which 
investors make their impact on the startup. 

We will first examine a key early decision — whether to self-ftind or to take money from outside 
investors. We then delve into the implications of taking money from three of the most common outside 
options: friends and family, angel investors, and venture capitalists. We will focus on four factors that 
differentiate these options: the investors' sources of capital, which shape their motivations and 
incentives; the founders' abilities to access the investors; the investors' potential to add value to the 
venture; and the costs and risks faced by founders who take capital from each particular type of 
investor. Throughout this chapter, we will focus on the decisions faced by founders and teams 
including Evan Williams and the founders of Ockham Technologies, a sales-automation software 
startup. 

SELF-FUNDING 

As with the decision to found solo or to build a founding team and with the decision whether or not to 
conplement the founding team with hires, founders have a "go it alone" option when it comes to 
investors. When Jim Triandiflou and Mike Meisenheimer started Ockham Technologies, they quickly 
attracted the attention of Monarch Capital Partners, which offered them $2 million. But they decided 
not to take the money and instead seeded the startup with $150,000 of their own money. "We just felt 
that we should go sell something [first]," Jim explained. "We knew we'd make the company more 
valuable by doing that and first getting some validation of our idea." Serial entrepreneur Frank 
Addante, after one startup failed and he had to give back VC fiinding, decided to self-fimd his next 
company, avoiding investors until he had a solid business plan and proof of concept. 

In 77% of founding teams in my dataset, at least one founder contributed seed capital early in the 
life of the startup. Even in those teams, though, the capital usually runs out pretty quickly. In the 
startups that had not yet raised an outside round of financing, the median monthly "burn rate" (the 
amount of cash being used) was $75,000 and the median startup was a little more than four months 



away from running out of cash. A startup with a small cash cushion is more vulnerable to liquidity 

problems and more likely to disband.^ As Professor Bill Sahlman points out, in startups, money buys 
time: time to experiment, to collect and evaluate data about what worked and what did not, and to 
adjust the strategy and operations based on what was learned. 

Barry Nails 's first startup, for example, was a completely self-fiinded two-person consulting 
business. Barry recalled how hard those years were: "There were two absolutely painful parts — I 
never got a day off and this was all my money. . . . When there's no payroll, it means you have no 
money [to pay yourself] either, so you're as broke as you can get. It was very tough financially and 
emotionally." When Barry ran out of money even for groceries, he said, "I started calling [customers], 
telling them I needed to collect from them I started working my customer network really hard." 
Eventually the pace of the business and lack of financing drove Barry back to the job he had left at 
GTE. 

The startup's industry, the macroeconomic context in which it operates, its degree of capital- 
intensity, and its business model all have powerM impacts on its financing strategy. ^ However, even 
when all of those factors point to trying to raise money from investors, some founders choose to avoid 
external sources of financing. Such founders are usually motivated to retain control of their startups, 
to be able to make all the decisions themselves, and not to have to spend time managing investors 
rather than growing the business. Either from prior experience or from the stories they have heard 
from mentors and from more experienced founders, they have learned the disadvantages of outside 
fiinding (which will be described below) and decided to avoid them Brian Scudamore launched his 
startup with the clear idea of never involving investors: "Even if someone would have offered, I 
wouldn't have taken outside money," he said. "It's something I got from my dad — that you don't go 
out and take money from someone. You should start small and grow it on your own."* 

It is possible to get away with this if the business can bring in customer revenues very early; for 
example, by consulting or by getting customer prepayments. For businesses with tangible assets, such 
as capital equipment or accounts receivable, which can serve as collateral, debt financing is an 
option."^ But high-potential startups in the technology and life sciences industries rarely have such 
assets, so debt financing is rarely an option for them. Among the startups in my dataset, only 2% 
raised debt financing, with this percentage remaining relatively consistent across all stages of 
conpany evolution. 

WHEN INVESTORS BECOME DESIRABLE 

Startups for which alternate sources of capital are not available, or whose growth ambitions require 
more capital, will need outside investors.^ Figure 9.1 shows the core questions that should be asked 
by founders and how the answers should lead founders to target different types of investors. As 
shown, founders have to assess whether they are still missing the human capital, social capital, and 
financial capital needed to grow the startup and to compete successfiilly; whether the different types 
of investors can play a role in filling those gaps; and whether the risks of involving those investors 
are worth the potential benefits of doing so. 



Startup capabilities and needs: 
Founder's background (Chs. 2-3) 
Cofounder strengths (Chs. 4-6) 
Hires' strengths (Ch. 8) 



Have important gaps in 
human capital or social 
capital? 



No 



1'" 



Need a lot of 
financial capital? 



No 



1'" 



Willing to risk 
losing control and 
to give up payoff 
from small exits? 



No 



Self-fund 
l^Yes 

Able to self-fund 
the startup? 



Target experienced 
angel investors or 
advisors 



Reassess capital 
needs or control/exit 
preferences 



No 



Target friends and 
family or inexperienced 
angel investors 
(then revisit needs) 



Yes 



Target venture capitalists 



Figure 9.1. Central "Investor Dilemmas" Questions for Startups That Can Attract Outside Financing 



Outside investors can provide founders with two broad benefits: (a) the capitals needed to get to 
the next stage of development and (b) improved governance, which includes bringing additional 

discipline to the startup,^ helping hone its strategy, and providing outside perspectives about the 
startup's prospects and its people's abilities. When founders consider potential sources of capital, 
they have a wide variety of options, each with its own pros and cons. The spectrum of investors is 
usually described as ranging from "dumb money" or "money-only" investors who only contribute 
financial capital, to "smart money" or "money-plus" investors who contribute all three types of 
capital. In between, there are many options; the "angel investor" category alone includes a wide 
spectrum of investors, ranging from inexperienced investors informally investing in the company to 
"super-angels" or to formal and experienced multi-angel groups that approximate early-stage VCs in 
their attention to due diligence and their relevant prior experience. We lack space here to cover all of 
the options, but below we examine the pros and cons of three prototypical investors: friends and 
family, angel investors, and venture capitalists. For each type of investor, Figure 9.2 summarizes the 
key differences, which are detailed in this chapter. 

These three types of investor tend to enter startups sequentially — first friends and family, then 
angel investors, and finally venture capitalists.^ However, the sequence should be seen as a sharply 
narrowing fiinnel, because many more startups can raise money from friends and family than from 
angels, and, in turn, many more can raise from angels than from VCs.* 

It should be noted, though, that many founders, even when they have started raising money from 
outside parties, continue to invest some of their own capital, often to maintain ownership stakes or to 
show confidence in the startup's prospects. As shown in Figure 9.3, even in their A-round of 



financing (the first round where they tap outside investors), 32% of founding teams continue to invest 
their own capital in the round, followed by 16% of founding teams who invest in the B-round and 

13%) in the C-round.^ However, once a startup's founders can no longer finance it themselves, they 
often look to the next easiest source of financing: their fi*iends and families. 



Imvslor's 









Vitlne Txpic^Hy Added 






Investor 


of C42f>ltill 


Atc fssinf; the Unvslor 


to Startup 


Riiks to Founder 


Friends 


• Pcfvonjl 


• F.isicst. Ptrional 


• I.itric to none. F.imily 


• 


.VLiy ^end an unprofessional si^al 


an J family 


wealth 


iKtwork% oi tore 


member* often do not 


• 


l.ax ^tandardt may enable fourulers to 


(snuU) 


founders arc tapped. 


have business skills. 




pursue marginal tdras 










* 


Pbying with Arc 


AngFl 


• Personal 


• Intermediate. Third- 


• May have broad 


• 


Less *constnKtivc discipline" and 


invcsion 


wealth 


part)' networking 


buiincss experience 




oversi^i/assistancc than VCs 




(medium 


and impcrwrul 


• Role on the foimdcr s 


• 


Can be difficult to manage 




to large) 


searches such as 


board 


• 


\lay introduce complications that later 




Angel forums can be 


• Social capital 




turn off po<cniial investors 






Uicd. 


• More tinancial capital 







than friends and family 



Venture 
capitalists 



Limited 
partners 
(very brgc) 







\\\:2!th 


( ..ritr,,l 


Flardest. .^ngel 


• More — and more 


• VCs demand 


* Board seats 


investors and 


predictable — financial 


equity, which 


• Protective 


professional contacts 


capital than angel 


dilutes founder 


provisions and 


are often used. 


investors 


stakes 


supermajoriry 




• Social capital (to tind 


• Ijquidation 


voting rights virith 




hires, future investors. 


preferences 


preferred stock 




and professional CEOs) 


affect exits 


• Differing risk 




• Rcputational effects 


• Forced vesting 


levels than 




• (juidance through 


of founder 


founders can 




taking an actiw role on 


equity stakes 


cause conflict 




the board and rcf^ular 


• Forced equity 


• Staging of funds 




communication 


reallocations 


• Drag-along rights 






among the 


can affect exit 






teams 





Figure 9.2. Pros and Cons of Investor Options 



90% 



80% 
70% 




60% - 
50% 

40% - 

30% - 
20% 

10% - 



0% - 


A-round 


B-round 


C-round 


D-round 


Founder Capital 


32% 


16% 


13% 


13% 


Angel Investment 


49% 


34% 


30% 


31% 


• VC Investment 


68% 


82% 


85% 


84% 


B Corp. Strategic Invt. 


12% 


19% 


27% 


29% 


— — Debt 


2% 


3% 


3% 


3% 



Figure 9.3. Types of Investors Participating in Each Round of Outside Financing in the CompStudy Dataset 

ACCESSIBLE BUT RISKY: MONEY FROM FRIENDS AND FAMILY 

Founders often have a small set of friends and relatives with whom they have frequent contact, long- 
term trusting relationships, and an emotional connection.^ Such prior relationships tend to make 
friends-and- family investors easier to access than angel investors, who also invest their own capital 
but tend to be more "professional" about it than friends and family. Investments from friends and 
family are often what make a startup possible in the first place. Meg Hourihan's parents invested in 
Blogger; Pandora used money from the father of a college friend of one of the cofounders; Scott Cook 
of Intuit, described below, took loans from his father. 

While early investments from friends and family can be crucial, they also have certain limitations 
conpared with more professional investments. Because friends and family are investing their own 
money, they usually invest much less than professional investors do. Because friends and family 
typically lack expertise or credibility in the most important business issues, they usually have little 
human or social capital to contribute. Because friends and family are usually inexperienced investors 
and not very hard to convince, their investments don't boost the startup's credibility in the eyes of 
outside parties such as potential customers or other potential investors. 

Friends and family are typically investing to support someone they love, like, or admire — not to 
make a lot of money. Lew Cirne, whom we will study in Chapter 10, raised $100,000 in seed money 
for his startup from his family and friends. "They didn't understand any of it," he explained, "but they 
were willing to bet on me." While such faith grants the founders a certain freedom from the demands 
and scrutiny that would come with professional investors, it also introduces a much more extreme risk 




if the startup fails, as discussed below. 

Motivation to Take Friends-and- Family Money: Comfort, Desperation, or Ignorance? 

The advantages and disadvantages of friends and family as investors are similar to their advantages 
and disadvantages as cofounders (discussed in Chapter 4). Such people are easy to reach, and there is 
an initial feeling of comfort and trust. Many founders jump at the chance to take this "easy money"; as 
one said, "Always take the money. You have to start somewhere." To the extent that founders tend to 
be very confident in their startup's prospects, they may even be more inclined to take money from 
friends and family, happily imagining those close to them sharing in the startup's fiiture success. One 
founder-CEO, acknowledging that complete certainty is almost never attained, said, "If you know that 
you won't lose their money, you should take it." In fact, whether a founder is willing to risk taking 
money from friends or family may be a good indicator of whether or not that founder has laid a solid 
groundwork for success. One founder-CEO said that "if you think there is a real possibility that you 
might lose their investment, maybe you haven't addressed all the risks" and should take the hint to 
think again and try harder to reduce them 

However, many other founders see friends and family as, at best, a last resort. Brian Scudamore 
was uncomfortable with the idea of taking money from relatives, even though at one point his startup 
was about to fold for lack of cash. He explained, "My dad could have given me the money. However, 
I didn't want him to worry that things weren't going well ... I wanted to show dad that I could do it 
myself, even if that meant stretching things a bit too far. Also, taking money from him would cause 
problems if things went sour." 

Brian's last point worries many other founders and investors. Just as cofounding with friends and 
family creates the Playing-with-Fire Gap discussed in Chapter 4, taking money from friends and 
family can put important — even crucial — relationships at risk and, if thing go badly, leave them in 
ruins. Founders who are confident in their startups' prospects and who are passionate about their 
ideas often discount the possibility that the startup could fail, but founders and investors who have 
already been burned by playing with fire take a very different view. One experienced advisor of 
startups observed, "If you go into business with money from friends and family, then you will either 
lose the business or lose your family, and have a very good chance of losing both." A serial 
entrepreneur explained, "Out of principle, I do not include family inbusiness/financial transactions or 
pursuits. The fact is, money stands to be the very thing that can dissolve the strongest of relationships. 
Added to the prospect of 'family drama' (of which everyone seems to have plenty), it's not worth it." 

One way around this can be to treat an investment from a friend or relative as a gift rather than an 
investment; that is, with no particular expectation of being paid back. As one founder put it, "To 
avoid any problems, just ask them to close their eyes and donate money to you." Speaking from the 
other side, one investor said about investing in a startup founded by a friend or relative, "I sometimes 
do it, but I regard it as a gift, not an investment." 

In addition to introducing playing-with-fire problems, taking money from friends and family may 
send a negative signal about the startup's prospects, especially if the startup is beyond the early 
launching stage. One serial entrepreneur observed that founders who fail to raise capital from 
professional investors and then have to raise from friends and family should think twice about the 
startup: "If you can't convince someone more objective than your friends and family to invest in you, 
there's probably a fiaw with your business, which will result in them losing their money and you 



feeling terrible. Even if you have a dozen rich uncles, I would suggest you find someone outside that 
circle who believes in you and your idea." An experienced investor said, "It's a shame to lose friends 
or family because they invested their savings in the business and lost it all. It becomes a tragedy when 
the business concept was flawed from the beginning. . . . Why do people take such a risk when there's 
plenty of other money out there? Because [friends-and-family] money is too easy to obtain, which 
encourages people to start marginal businesses. Beyond that, friends and family do little diligence, 
have no real objective judgment, and can't advise/prevent you against doing something completely 
stupid."io 

Another experienced VC suggests that some founders may simply not know what their fiinding 
options really are: "The reason people usually go to [family and friends] is that they don't have other 
choices — or more likely, they don't know they have other choices." 

Pulling together these considerations, one serial entrepreneur offered this rule of thumb: "If your 
family comes from a business background and are well-informed about finances and they know to 
keep business and friendship strictly apart, you should utilize them. However, there are clearly too 
many ifs, ands, or buts in this scenario, so founders would be well advised to search for financing in 
other places." Just as cofounding with friends and family should be accompanied by carefiilly 
constructed firewalls and explicit discussions about worst-case scenarios, so too should any 
financing from friends and family, or else founders should resist the temptation to take such money. 

"Burning the Boats": Productive Motivation or "Entrepreneurial Suicide"? 

Taking money from friends and family may not only make a startup possible, but also profoundly 
affect how — and even why — the founders try to grow or exit their businesses. For instance, taking 
such money may enable founders to persist through the scary drops of the entrepreneurial roller 
coaster, not simply because they have money with which to do so but because they don't dare give up. 
Scott Cook, the founder-CEO of personal-finance software company Intuit, tried to avoid raising 
capital from friends and family. But 25 failed pitches to professional investors later, he gave in 
and borrowed from his parents' retirement savings and from his cofounder's friends. Altogether, he 
"spent about $350,000 on Intuit, a sum pieced together from life savings, home-equity credit, credit 
cards, and loans from his father." During one particularly difficult period, he said, "What kept me 
going was just fear that I didn't know how I'd ever pay back the money." 

Similarly, Tim Westergren, founder of Pandora Radio, persisted with his startup long after others 
might have called it quits, in large part because he could not face the idea of losing the money 
invested by his cofounders' friends or reneging on the $1 million in salary deferrals the company 
owed its employees, most of them personal friends of the founders. As Tim explained, "I've brought 
everyone into it and can't turn back. My persistence might look like a virtue, but I have to put myself 
through this because of all the people I've involved who I have ties to." Indeed, objective observers 
might question whether such persistence is virtue or vice, skewing founders' incentives and possibly 
leading them to throw away years of their lives, or to throw good money after bad. 

Some founders refer to taking money from friends and family as "burning the boats," referring to 
the legend of the Spanish conquistador Hernando Cortez, who, after arriving in South America with 
700 men seeking to conquer the continent for Spain, ordered them to burn their boats so they could not 
even think of retreat. One founder said, "Once the money is in, a whole new dynamic is involved. 
With the relationship at risk, the entrepreneur (the good ones anyway) will move heaven and earth to 



not lose the investment. And moving heaven and earth is often what it takes." 

An experienced startup advisor takes a more negative view, comparing "burning the boats" to 
"entrepreneurial suicide." He explains, "If one sets oneself up in situations where you must 'succeed 
or die,' then your fundamental motivation must be called into question. . . . Instead of taking prudent 
risks to maximize your chance for succeeding, you live constantly on the edge in a world of 'possible 
success.' If this is the case, then the business is your 'addiction' that you're asking your fi-iends and 
family to support. Further, founders and entrepreneurs should have a 'safe haven' if they are to have 
any hope to survive the rigors of the startup company lifestyle. Where do you go for physical/psychic 
nourishment and comfort when the world is crashing in on you? Family serves this purpose often and 
best. 'Burning your boats' undermines this support structure and thus is setting oneself up for failure." 

For other founders, taking money from friends and family may make them more risk-averse than 
they might otherwise have been, which can result in a lower exit value. An experienced investor 
explained, "When an entrepreneur has all of their own personal wealth or their family's wealth tied 
up in a deal, they are more likely to be risk-averse and more likely to sell out at a smaller number." 
Founders who prefer a higher-probability small exit to a lower-probability large exit would see this 
as a good thing. 

Although the effect of taking money from friends and family will differ depending on the founders 
involved and the specifics of the situation, friends-and- family investors infroduce a more emotional 
element to the business as it moves forward, regardless of whether that heighted emotion is beneficial 
or defrimental to the founder and the startup. 

fflGH VARIANCE: ANGEL INVESTORS 

"Angel investors" refers to a wide range of individual investors who invest their own money and 
usually don't already know the founder. They invest at an earlier stage of startup development than do 
venture capitalists, often with the goal of atfracting VCs to invest in a subsequent round of financing. 
One analyst of angel investments writes, "The angel market is essentially the farm system for the next 
wave of high growth investments."^^ 

Angels differ from friends and family in almost all of the dimensions described above: 

• Motivations — Angels are much more financially driven than are friends and family. They 
typically prioritize getting a good financial return from their investments, but many also have 
nonfmancial motivations, such as mentoring the next generation of founders, passing on their 
own hard-earned lessons, and applying their deep industry knowledge to help build the next 
generation of companies. 

• Access — ^As shown in Figure 9.4, 13% of angels were investors in a founder's previous 
startup. However, for the typical first-time founder, angels are much harder to reach than are 
friends-and- family investors and there is far less initial trust and comfort. Most founders have 
to rely on weaker ties to access angels, either being introduced by someone they know or 
applying to make presentations to "angel forums" such as TechCoast Angels, a Southern 
California network with over 300 angels, or CommonAngels, a Boston network of about 70 
angels. In my data, 58% of startups were referred to their eventual angel investors by 
someone with a common tie (i.e., by a friend or family member of a founder or of an early 
executive in the startup), 1% of the startups approached the angel by submitting a business 



plan or making a cold call, and 4% were referred to an angel by a service provider such as an 
accountant or lawyer. 

Human capital — Occasionally, a friend or family member has experience or contacts that are 
directly relevant to the startup. For instance, if the person worked in the same industry or had 
worked in another startup, he or she might be able to provide advice and guidance or be able 
to introduce the founders to potential customers or partners. However, such value addition is 
more likely to be gained from professional angels, who usually have more business 
experience. Early on, some play an active counseling role and may even sit on the board of 
directors until VCs begin investing. But their participation on the board usually does not last 
beyond the entry of VCs, and thus is less sustained than their participation in providing 
capital; by the second round of financing, few angels still sit on startup boards. 






Initial Angel inve$tor(s) 



Initial VC investor(s) 



□Other source 



35% 



□Submit BP/cold call 
□ Investorin founder's prior startup 



21% 



18% 



□ Other eNsc's friends & family 



7% 



I CEO's friends & family 



19% 



Figure 9.4. Sources of Leads to Initial Angel and Venture-Capital Investors 



• Social capital and credibility — Depending on the backgrounds and reputations of its angels, a 
startup may gain a lot of credibility from them.* Angels can also play a direct role in bringing 
VCs into a startup. In 10% of the startups in my dataset, angels introduced the startup to a VC 
who eventually invested in it. The angel may also have customer contacts, industry expertise, 
or relevant skills that can add a lot of value to the startup. 

• Financial capital: Amount invested — Unless the founder chose his or her parents (or social 
circle) very carefiilly, friends and family are usually limited in the amount of money they can 
invest in the startup. Angels can fill the gap between friends-and- family investors and 
institutionalized investors such as venture capitalists. Between 2001 and 2009, the average 
size of an angel round was approximately $450,000, which falls right in between the amounts 

invested by friends-and- family investors and by venture capitalists.^^ Although most 



individual angels invest modest amounts, some organized angel groups invest much more; 
TechCoast Angels targets investments of up to $ 1 million, and CommonAngels targets deals 
that total $500,000 to $5 million. 

• Financial capital: Continued investment — Although angels mainly get involved in the 
earliest round of outside financing, many also try to invest in later rounds in order to maintain 
a minimal level of ownership in the startup. (This is also in contrast to friends-and-family 
investors, who ofl;en do not continue investing substantial amounts in later rounds.) As shown 
in Figure 9.3, angels participated consistently in 30% to 34% of the second through fourth 
rounds. 

• Personal risks — ^Although, as we will see, taking money from angels can increase some types 
of risk dramatically, the playing-with-fire risk is much lower with them. There is usually no 
prior social relationship that will be damaged by failure, and professional angels are usually 
investing only a small percentage of their net worth in any one startup. 

Among professional investors, venture capitalists usually get the lion's share of attention, but angel 
investors actually account for many more investments per year. Analyzing data from the University of 
New Hampshire Center for Venture Research's annual surveys of angel investors, I found that angels 
accounted for an average of 49,000 investments a year between 2001 and 2009, an order of 
magnitude greater than the annual number of VC investments during the same time period.* But VC 
investments, as described below, tend to be much bigger; despite the much greater number of angel 
investments, total angel capital invested per year averaged $22 billion, less than the venture capital 
invested annually over the decade. 

One advantage of angel investors is the diversity of industries in which they invest. Venture 
capitalists tend to target technology and life sciences startups; for instance, "Internet-specific" 
investments accounted for more than 20% of capital invested by VCs in half of the years between 

2000 and 2009 and for over 30% in 2000 and 2008.^*^ Startups in other industries have a better 
chance of being targeted by angel investors. Excluding the temporary spike in software investments 
during the dot-com boom, no single sector accounted for more than 20% of the capital invested by 
angel investors in any year between 2001 and 2009.^^ 

The downsides of many angels include a general lack of understanding and experience with respect 
to both the business and their role as investors. In addition, because angels each invest smaller 
amounts, founders end up with a larger group of investors to manage. As Dick Costolo of FeedBurner 
explained, "We didn't know any VCs but we knew people who had sold companies and were 
interested in investing. We got a dozen accredited investors who took all common stock, no preferred 
stock at all. So far, so good. However, there were a few investors in the syndicate we didn't know 
well and they became problematic. They were making high-risk investments in a new company, but 
didn't feel like they could afford to lose their entire investment. It was a recipe for disaster. I had 13 
people who, now that they had $20,000 invested, wanted to call me and ask about an article in the 
Wall Street Journal, taking 45 minutes of the CEO's time when he should be running the business!" 

Dick also explained that angel investors do not demand the level of "constructive discipline" and 
accountability that more professional investors would require of a startup: "We didn't form a board 
because it was an angel round. We did things randomly for six months, paying ourselves randomly. 
We ended up having to go back and put our books in order when we did a [VC] deal later and it was 
a painful process. A board would have forced us to do things right and would have pointed out that 



the way we were hiring people could have used more rigor." 

At Ockham, Jim Triandiflou's experience negotiating with angel investors illustrates some of their 
advantages and drawbacks. When Jim was ready to raise funding for Ockham, he first approached a 
group of angel investors led by Bobby Crews, a real estate developer from Texas whom he had met 
through a business associate. "Their term sheet was extremely interesting," Jim recounted. "They 
would give us $10 million: $3 million now and another $7 million as we needed it. They would take 
50% of the company on Day One and then would keep their hands out of the business." Jim also liked 
the fact that, if he took the Texas money, Ockham's board would be small, and the founders would 
have two of the three seats. The major drawback, explained Jim, was the angels' lack of industry 
experience: "They didn't understand anything about the business. Bobby's a real estate guy! He would 
talk about it at cocktail parties and it'd be kind of cool, the little Atlanta software company he's 
funding." The lure of "no strings attached" angel money was quite powerful for Jim, as it is for many 
founders. 

STRIKEOUT OR HOME RUN: VENTURE CAPITALISTS 

Venture capitalists are professional investors who focus full-time on investing in high-potential 
startups. They receive business plans from the founders of young startups,* evaluate the plans, meet 
with the founders, perform due diligence to investigate the startup's team and potential, negotiate 
investment terms with the startups in which they want to invest, then help build the startup (often as 
board members) in the hopes of exiting from their investment at an appropriate time. 

VCs raise capital from limited partners (LPs), most of which are large institutions — foundations, 
university endowments, public pensions, and the like — that invest in a wide portfolio of assets and 
allocate a relatively small percentage of their capital to venture capital. The VCs have a fiduciary 
duty to those LPs and are evaluated on the financial returns from their investments. VCs are paid a 
management fee (usually 2%-2.5% of their assets), which they use for salaries and operating 
expenses, and a share of the financial gains ("carry") that averages 20% of the aggregate profits from 
their investments. VC firms are usually founded by a small group of general partners (GPs) who have 
to raise the funds they will invest, organize the operations of the firm, and sometimes hire junior 
people (analysts, associates, principals, and others) to whom they can delegate tasks in order to make 
better use of their own time. 

More than any other type of investor examined here, VCs are motivated to prioritize financial gains 
and take any actions they believe will increase their returns. VCs raise funds that last a decade (often 
with an option for short extensions); they usually invest the majority of the fund during the early years, 
then have a few years to build and exit from each startup before the fund ends. VCs who want to 
continue investing must therefore raise a new fund every few years, requiring them to justify their 
performance and the solidity of their strategy to LPs. To maximize financial returns while reducing 
the risk of their investments, VCs seek to invest in a portfolio of startups, often numbering in the 
dozens for each fund. For each investment, VCs are seeking to return a large multiple of the amount of 
money they have invested, so they look for startups with the potential to be extremely big. The most 
common exits for VC investments are acquisitions by larger companies (as when Google bought 
FeedBurner), initial public offerings (as when Google itself went public), or, much less attractive, the 
bankruptcy of the startup. In a recent study of more than 22,000 venture-backed companies from 1987 
to 2008, 26%) of the startups had merged or been acquired, 9% had gone public, 15% had been 



liquidated or had gone bankrupt, 19% were expected not to return any money to shareholders, and 
31% remained private. 

In the following two sections, we will see that founders can gain a lot, both financially and 
nonfinancially, by taking VC money. But there can also be serious losses of ownership and control. 



Benefits of Venture Capital 

Venture capitalists can provide founders with more financial capital, social capital, and human 
capital than the typical fi-iend-or-family investor or angel investor can. Although there are always 
some who question whether those benefits outweigh the costs of taking VC money, many 
entrepreneurs persist in seeking VC investment. In fact, the more reputable, experienced, and well 
connected the VC firm, the more unfavorable the terms an entrepreneur will accept in order to secure 
an investment from it: A study of entrepreneurs who had received offers from multiple VC firms 
showed that, on average, entrepreneurs were willing to give up more than $4 million in valuation to 
get a better VC firm on board.^^ However, it is important to note that the entrepreneurs in this and 
similar studies are those who chose to seek VC investment to begin with; some of those founders who 
did not seek VC investment are likely to disagree that a VC investment is worth that price, as we will 
see below. (Capturing the spirit of this latter perspective, one founder insisted, "The goal of every 
board meeting is to end it.") 



Financial Capital 

VCs are attracted to high-risk/high-reward investments, which excludes most startups and which 
makes VC portfolios more volatile than portfolios of lower-risk/lower-reward investments would be. 
Within each portfolio, VCs tend to have a few "home runs" they hope will make up for the majority 
that either fail or produce only small returns. One study estimated that, if the financial market's 
overall volatility is defined as 1.0, VC returns have a much higher conparative volatility (called a 
"beta") of 1.7.25 




Figure 9.5. Capital Invested and Pre -Money Valuation for Each Round of Financing 



As shown in Figure 9.5, even the smallest round that includes VCs (the A-round, which averages 
$3 million) is far bigger than rounds raised from friends-and-family investors and angel investors, 
which rarely surpass $1 million. Subsequent rounds get even bigger, especially in life sciences, 
where startups are often more capital-intensive, requiring nearly $10 million per round after the 
initial round. 

VC investments are generally not only larger than friend-and- family and angel investments, but also 
longer-term. VCs invest with the intention of continuing to invest through multiple rounds of financing 
and usually budget their capital careftilly to make sure they can continue to invest if the startup is 
making good progress (as described below), in contrast to friend-and-farrdly investors who rarely 
continue into later rounds and to angels whose continuation is uncertain. Each round of financing, VCs 
decide whether to continue investing in each startup or to shift their capital toward more promising 
startups. 

At the same time, the availability of professional capital is also heavily dependent on market 
conditions. During the dot-com bust, total VC capital invested plunged by 52% between 2000 and 

2001 and by another 41% in 2002.^^ VC investments then jumped 57% during the turnaround in 2005- 
2006, only to drop again by 37%) during the early part of the recession in 2008-2009. Even if VCs 
invest with a long-term perspective, their sensitivity to the business-cycle roller-coaster ride is quite 
pronounced, limiting the confidence that startups can place in venture capital as a consistent source of 
fimding. 

Social Capital: Connections and Credibility 

Founders of startups often rely on their investors and board members to make up for their own 
potentially fatal lack of connections, particularly connections to those who might serve as fijrther 
sources of capital. In a survey of VCs regarding the startup-related activities in which they were 
most involved, they ranked obtaining alternate sources of equity financing and interfacing with other 

investors second and third, respectively.^^ (Their top-ranked activity is discussed below.) In a 
follow-up survey asking startup CEOs to rank the contributions made by their VC investors, 
introductions to other potential investors was ranked second.^^ In my own dataset, VCs were more 
effective than other investors at making connections to ftiture investors. For instance, initial VC 
investors connected 31%o of startups to subsequent VC investors, while angel investors did so for 
only 10% of the startups. Similarly, VC investors connected 14% of startups to their initial corporate 
investors;^^ angels did so for only 5% of the startups. 

High-status investors can also give a young venture some of the legitimacy it lacks. Startups with 
more prominent investors receive higher public-market valuations; the boost from prominent 
investors is especially significant for younger, less-proven startups for whom a boost in credibility is 

most important.^ ^ Frank Addante experienced the "halo efifecf of top-tier venture firms when he 
began talks with Sequoia, a famous Silicon Valley VC firm Although Sequoia's term sheet for 
Frank's startup, StrongMail, was not as attractive as some of the other offers he had in hand, Frank 
chose to do business with Sequoia because of their legendary reputation for helping build valuable 
startups. After meeting with Sequoia's partners at their oflRce, he recalled, "I realized I was sitting in 
the same chair Steve Jobs had sat in, the same chair Larry Ellison and a lot of other big names had sat 
in." Having Sequoia on board helped convince potential hires and customers of StrongMail's 
prospects. As we saw in Figure 8.1, by the B-round of financing, investors are the second biggest 



source of executive hires, accounting for 19% of hires and helping make up for the founder-CEO's 
declining (though still strong) centrality in hiring. 

Human Capital: Guidance and Mentoring, Both Informally and through the Board of Directors 

Do VCs provide guidance to their startups? In a survey of VCs to gauge the startup-related activities 
in which they were most involved, the highest-ranked item was "serving as a sounding board to the 
[executive team]."^^ Through their own websites, interviews, blog posts, and other forums, VCs tout 
their ability to add value. For their part, founders who have taken money from VCs do report that their 
investors had an impact, often positive, on how they ran the business. In the aforementioned survey 
asking CEOs to rate their VCs' contributions, the top-rated activity was serving as a sounding board 
to the executive team.^^ In another study, startups that received VC financing were 47.7% more likely 
than those with other kinds of investors to report that their investors were influential in shaping human 
resource management policies and initiating other professionalized practices.-^'* In yet another study, 
VC firms with a high level of industry experience were able to get their term sheets accepted by 
entrepreneurs despite offering valuations that were, on average, 10% lower than those of competing 
offers. -^^ This seems to indicate that founders expect to benefit more from the more experienced VC 
firms. Founders should recognize, though, that even if the VC firm has a great reputation, that is no 
guarantee that all of the firm's individual partners will be effective and committed. As a general 
partner from a top firm observed, "Not all money from Sequoia is the same color — it's different if 
you get [senior partners] Mike Moritz or Doug Leone instead of one of their newbies." 

How is this valuable guidance being delivered? A lot happens informally or during one-on-one 
meetings between investors and founders. The issues covered in those discussions can vary greatly 
from startup to startup and from investor to investor, but are most likely to include strategic, 
personnel, operational, and financing issues. For instance, a VC will often serve as a "strategic 
counselor" who helps the founder-CEO hone in on the central strategic uncertainties facing the startup 
and who challenges the founder's core assumptions, "generally, pushing the CEO to step back from 
the day-to-day tactical operational activities and focus on the true value- creating decisions that need 

to be made."-^^ The frequency of these informal meetings tends to change as the startup matures. As 
one VC explained, "The stage of the business has the biggest effect on mentoring frequency. In the 
first two years, as the team is being built, the strategy is getting set, etc., [the interactions are] many 
times per week. Some email traffic, some instant messages, some calls, some in person one-on-ones." 
After that, as the team and strategy solidify, such interactions become less frequent and more ad- hoc. 

Another vehicle by which VCs provide guidance is formal meetings of the startup's board of 
directors. In fact, the very existence of a formal board is often a direct outcome of VC involvement. 
As described in Chapter 5, "Role Dilemmas," startup boards usually begin as informal decision- 
making bodies made up of the founders and sometimes one or two trusted advisors. By the time the 
first round of VC financing closes, there is a formal board of directors that includes at least one VC 
investor. (When multiple VC firms participate in the round, either multiple VCs join the board or the 
VC who "led" the round does.) 

Thus, VCs not only demand a board, they generally demand to be on it, and the VC presence tends 
to increase over time. Furthermore, it tends to increase at the expense of founder representation. This 
is, of course, a significant control issue, as discussed below, but for our current purposes, it also has 
implications for guidance and mentoring, for the following reasons: 



• Longevity. In contrast to angels, who usually leave the board of directors once VCs begin 
investing, VCs tend to remain on startup boards much longer. This longevity can help them 
provide better guidance than if board members were regularly being replaced, and CEO- 
board relationships are often more productive when the CEO has a chance to develop 
relationships with the board members. 

• Size. The larger the board, the more industry knowledge and business expertise the CEO might 
be able to tap. In my data-set, after the first round of financing, 80% of boards had 3 to 6 
members, with an average of 3.9. After the second round of financing, that average was 4.8 
members. After that, the growth slowed; the average was 5.2 members after both the third and 
fourth rounds. By the fifth round of financing, 89% of boards had between 5 and 8 members, 
averaging 5.7.* (Despite their potential advantages, though, larger boards also introduce their 
own challenges, described below.) 

• Experience. Board members with significantly more experience can gain added authority — 
the more so the smaller the board. The VC members of a startup board are likely to have more 
pattern recognition than the founder members and, in particular, are more likely to have 
experience as board members. In my dataset, 48%) of outside directors had prior experience 
as directors. Most of those were venture capitalists who had sat on the boards of other 
startups in which they had invested, had seen common problems encountered by management 
teams, and had negotiated multiple financing rounds and company exits, in contrast to the 
founders themselves, who accumulate few such experiences. 

• Frequency of meeting. Startup boards also meet more frequently than the quarterly schedule 
maintained by most large-company boards, although the frequency itself changes over time. 
This reflects the fact that startups are evolving with a speed and intensity not found in large 
corporations. It also reflects the fact that startups are often being run by an executive team 
with less experience than would be found in a large company. Such a team may need a 
different quantity and quality of guidance from its board, guidance that becomes available 
once the board includes VC members. In my dataset, boards averaged 7.9 meetings a year, far 
more often than large-company boards. There was a wide variation in meeting frequency, but 
a consistent pattern was that boards met more frequently when the startup was younger, then 
less frequently as the startup matured. An experienced VC talked about how he insists on 
monthly board meetings while the startup is young, but that "[o]nce the business is well 
'metric-ed,' is firing on all cylinders, and has a complete team, 60 days between meetings 
works." In other words, as ventures become more formalized and have metrics with which 
performance can be tracked, and as the executive team is conpleted, there is less need for the 
VCs to use board meetings to monitor the startup's performance and less need for them to 
provide guidance and recruiting assistance, so meetings can be held less frequently. 

All of these factors suggest that startup boards serve as powerfiil conduits for the influence of VCs, 
an influence that seems, on the whole, to be viewed from both sides as helpfiil to the startups. My 
research with Warren Boeker suggests another dimension of this influence; namely, that the CEO's 
background is strongly linked to the degree to which VCs are actively involved in the startup.* To 
examine this, Boeker and I used my dataset of 450 technology startups to examine the proposition that 
three aspects of the CEO's background — fiinctional background, years of work experience, and 
whether or not he or she was a founder — affect three aspects of the board — size, frequency of 

meetings, and composition (particularly the percentage of directors with prior board experience). 



We found that ditferent kinds of CEOs do indeed have very different boards. * These differences 
could affect how much value the board will provide and how challenging the board is to manage. 

First, CEOs with a midrange amount of prior work experience have the largest boards and meet 
most frequently with them. Both results suggest that those CEOs have the most opportunity to receive 
guidance from their boards. While this is good news for those CEOs with a moderate amount of 
experience, it does not bode well for inexperienced CEOs, who are most in need of guidance. At 
Odeo, Evan Williams struggled to develop a product roadmap and looked to his board for guidance 
about how to do so, but grew frustrated with how little usefiil advice he got from them 

Second, regarding fiinctional backgrounds, CEOs with experience in an external or output fiinction, 

such as sales, have smaller boards that meet less frequently and have less-experienced directors; 
CEOs with experience in an internal or throughput fiinction, such as technology, have boards that meet 
with them much more frequently. For instance. Wily Technology's founder-CEO, Lew Cirne, who had 
a deep technical background, would be more likely to meet frequently with his board than would Jim 
Triandiflou of Ockham Technologies, who had a deep background in sales. Apparently, for technical 
founders, getting the guidance you need from your board can also come at the cost of having a larger 
board to manage and having to spend more time meeting with it. 

Third, even afier controlling for differences in the CEOs' levels of experience and fimctional 
backgrounds, there was a very strong relationship between a CEO's founder/nonfounder status and 
the board's experience: Founder-CEOs had much more experienced boards than did nonfounder- 
CEOs. There may be several reasons for this. To the extent that founders are overconfident and 
passionate, and thus can benefit from having a board that can serve as a reality check on their 
tendencies, their more-experienced boards are more capable of playing that counterbalancing role. As 
Bagley and Dauchy suggest, "The most effective boards give independent, informed advice to 
management and challenge the CEO, rather than acting as a rubber stanp."^^ Ideally, says one VC, 
"Boards should be encouraging the type of learning to allow a company to 'pivot' by making 
important changes in its strategy." To the extent that some founders have trouble focusing on a single 
idea rather than pursuing new projects, boards can also serve as a check on that tendency. Evan 
Williams, for example, was famous for jumping from project to project and failing to focus on the 
core deliverable at hand. Evan himself recalled that, in his first startup, a family-fimded Internet 
publishing company, "I realized I had started 32 different projects during the previous year and not 
finished any of them I was constantly thinking up new ideas and found it better to abandon the current 
thing so I could work on the new idea. I was very shortsighted." Evan eventually closed the company, 
losing all of his father's investment. A strong board might have kept Evan on track and avoided this 
fate. 

All told, these results suggest that boards are often built with the CEO's weaknesses and strengths 
in mind. Except for the least-experienced CEOs — who seem to get less guidance than they might need, 
possibly because they are less attractive mentees — the CEOs who have the most need for guidance 
(those with a moderate amount of experience) seem likelier to get it. There is a wide-open 
opportunity for research that can show what is actually happening and why. For example, are 
founders seeking board guidance, or are VCs on boards making sure CEOs get it? 

I should mention a fiirther factor that can limit the intensity of VC influence through the board. The 
geographic proximity of a startup's VCs plays an important role in board composition and 
significantiy affects the degree to which VCs can engage in close monitoring and CEO counseling, 
which are most effective in person. Controlling for a wide range of differences among VC 



investments, Professor Josh Lerner found that the closer a startup is to its VC's office, the more likely 
the VC will be to take a seat on the startup's board. The probability that a VC investor located within 
five miles of the startup will serve on the startup's board is 47%; for a VC located over 500 miles 
away, the probability shrinks to 22%.^^ 

Costs of Venture Capital: Giving Up Ownership 

There are two major rights affected by ownership in a startup: economic rights and control rights. By 
selling equity stakes to investors, founders give up some of both. 

Changes in Overall Ownership 

In each round of financing, the founders/insiders sell more of their startup's equity to investors, 
decreasing their own ownership percentages. When the startup returns money to its owners — often 
through an exit when it is acquired or goes public but sometimes also from a share of the ongoing 
profits — the owners receive that money in proportion to their ownership, unless there are other 
investment terms that affect the payments, such as the liquidation preferences described below (and 
many other terms that we lack space to cover here). 

Figure 9.6, based on my dataset, shows how the percentage owned by each major type of 
shareholder changed after each round of financing. The most striking pattern is that after the C-round 
of financing, VCs finally owned more than half the equity — 53% on average. (After the B-round, VCs 
averaged 46%) ownership.) The increases in VC ownership were accompanied by steady decreases in 
the percentage owned by founders and employees, whose collective stakes plummeted from 41%) 
after the A-round to less than half that after the D-round. 

These changes do differ by industry. Life sciences startups tend to be more capital-intensive, 
requiring the founders to give up more equity in order to get the capital they need. After each round of 
financing, life sciences founders/insiders owned three to six percentage points less equity than did 
their technology counterparts. In addition, within each industry, the more attractive the startup and the 
fewer its weaknesses, the higher the ownership stake typically retained by the founders and inside 
executives.^^ 

Founders/insiders are particularly likely to lose ownership in a down round, commonly defined as 

a round in which the venture's share price is less than it was in the previous round, which means 
that the founders and insiders now have to give up more ownership to raise a given amount of capital. 
Across all rounds in my dataset, 1% were down rounds. Sometimes, a down round occurs when the 
startup misses milestones and underperforms. One startup, having earlier raised an initial round of $1 
million, wanted to raise a second round of $1 million. Because it had become evident that the startup 
was not going to achieve its milestones for bringing its system into operation, the investors demanded 
a higher percentage of ownership — and a lower share price — in return for their investment. As the 
company continued to experience "missteps and missed deadlines," the startup's investors invested 
another $2.7 million, bringing their ownership stake to 75% of the company's equity as the startup's 
price per share sank. 




0% 


A-round 


B-round 


C-round 


D-round 


E-round 


□ Other 


12% 


10% 


14% 


13% 


16% 


DVCs 


34% 


46% 


53% 


51% 


54% 


■Angels 


13% 


13% 


12% 


16% 


13% 


■Current founders & 
employees 


41% 


30% 


22% 


20% 


18% 



Figure 9.6. Capitalization Tables: Average Equity Holdings of Each Major Party 



Figure 9.7 shows that down rounds are also closely tied to the business cycle: During the high 
points in the cycle, such as the late 1990s and the mid-2000s, the percentage of down rounds 
decreases; during bust times, such as the early 2000s and the late 2000s, the percentage increases 
markedly.* Thus, a startup's exposure to the negative effects of a down round can be due both to its 
own ability to execute and to factors outside of its control. 



14% 



12% 



10% 




I I ™ I I I I I I I I I I I I > 
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 

Figure 9.7. Percentage of Financing Rounds That Were Down Rounds, by Year the Round Was Raised 



Liquidation Preferences: Giving Up "Small" Exits 



The percentage of equity owned by each party tells only part of the ownership story Other terms in 
the investment agreement affect how much each party would receive in various exit scenarios. 
Perhaps the most important of these terms affects each party's payouts, particularly when there is a 
small exit. That term is called the liquidation preference, which dictates that if the startup liquidates 
or is acquired, then the VCs get back the amount they invested, or a multiple of that amount, before 
other shareholders (most centrally here, the founders) receive any proceeds. "^"^ During the first round 
of financing, 78% of the startups in my dataset agreed to give their investors a "Ix" liquidation 
preference; that is, if the startup would sell for the amount of capital invested by those investors (or 
less), those investors would receive all of those proceeds and the founders and other early 
shareholders who own common stock would receive nothing.* (Once the liquidation preference has 
been paid out, the common stock holders usually begin receiving some proceeds. )'^^ 

Far more drastic are the liquidation terms to which the other 22% of startups agreed: 9% agreed to 
liquidation preferences of l.lx-2x, 5% agreed to 2.1x-3x^ and 8%) agreed to greater than 3x. In this 
most extreme latter category, the startup would have to exit for at least three times the amount 
invested in order for the founders and common-stock holders to see a penny of return. As startups 
raised fiirther rounds, the percentage of startups agreeing to preferences greater than 1 x increased, 
from 22% during the first round to 33% during the fourth round of financing. 

Investors require such terms so they can reduce their risks and align the founder's incentives with 
their own. Without such terms, the investors might be unwilling or unable to invest at all, would have 
to reduce the amount of capital, or would have to use other — more costly — terms. Why, though, do 
founders agree to such terms? Sometimes, it is simply because they sorely need the capital and lack 
the negotiating power to resist. However, another important factor is entrepreneurial optimism. Filled 
with passion for their idea, confident in their abilities, and dreaming of building a large and valuable 
company, founders often shrug off the possibility that their startup may sell for only a small "nice" 
price. They figure, "There's little chance we'll sell for less than $X, so if the VCs want a liquidation 
preference for that amount, I'll give it to them and I'll take the more likely upside of a big sale or 
IPO." At Lynx Solutions, the founders agreed to standard Ix liquidation preferences that meant that 
the startup would have to sell for more than $29.5 million (the amount invested by the VCs) for the 
founders to earn a dollar from the sale. Amid a market downturn. Lynx did not receive any offers 
greater than $25 million, which introduced major misalignment between the founders (who did not 
want to sell and go entirely unrewarded for their years of hard work) and the investors (who 
preferred a moderate exit to what they saw as a good chance that the startup would fail).* 

Some founders, having gotten burned signing such terms, begin to feel that their investors took 
advantage of their early passion by having them sign large liquidation preferences. They regret that, 
with all their eggs in one basket, they did not push back against the much more diversified venture 
capitalists rather than bear the risk themselves. But a wiser course would be for founders to take the 
small-exit scenario seriously from the beginning, however confident they are of something better. In 
fact, the more confident they are, the more helpfiil it would be to get some impartial advice on the 
rational probability of a small exit and consider whether they really are willing to give up any 
possible gains from such an exit. 

But aside from optimism and overconfidence, founders may have more rational reasons to sign 
such terms. Evan Williams, for example, agreed to a 1.5x liquidation preference for Odeo's 
investors, largely because he felt that time was running out for the startup to bring its product to 
market ahead of well-fimded competitors. Other founders think careftilly about the probabilities and 



seek outside advice to counterbalance their confidence and passion, then still decide that the increase 
in potential return is worth the new risks. 

Whatever the motivation, giving investors a liquidation preference increases the pressure on 
founders to "swing for the fences" and to adopt higher-risk/higher-return strategies, since they now 
have nothing to gain from a "small" exit to which they might otherwise have agreed. Founders who 
are incented to swing for the fences are more aligned with investors who likewise want to hit a home 
run, but founders (for whom this might be their only swing at the ball) do so at the cost of increasing 
their risk, often dramatically.* 

Investors Forcing Ownership Shifts within the Founding Team 

Investors may also have a powerfiil impact on ownership within the founding team itself The most 
common impact is when VCs insist that the equity held by founders and nonfounding executives must 
vest — that is, be earned back — over a preset number of years or after the achievement of 
predetermined milestones, in order to provide "golden handcuffs" to keep them working for the 
startup. (See Chapter 6 for details on founder vesting and Chapter 8 for non- founding executive 
vesting.) For instance, at an early Web startup, the venture capitalist insisted, as a condition of the A- 
round and as a way to incent the founders to keep working at the startup, that the two founders' frill 
equity stakes had to vest, effectively returning their economic ownership to zero and forcing them to 
re-earn their old stakes. Taken aback but needing the capital, the founders fought the change — in vain. 

A less common, but still very important, way in which investors can affect founders' stakes is 
when the investors insist on reallocating those stakes to more closely match the individual founders' 
contributions. Investors want the most important members of the team to own the largest equity stakes 
in order to both retain and motivate those members. Investors will therefore assess whether the 
"righf founders have the largest stakes and, if they detect too great an imbalance, will ofren insist that 
the founders reallocate their equity as a condition of the frnancing round. One founder said, "I was the 
cofounder of a startup and as part of the Series A, the investors actually required my stake to be 
increased. They felt the original founder's equity was lopsided and wanted to ensure incentives were 
properly disfributed across the team." A VC explained why he might insist on an equity adjustment 
before investing: "We spend time during our diligence process understanding how [the founders' 
equity splits] came to be. By the time we look at a company for investment, there has usually been 
enough development so that issues resulting from 'quick handshakes' or disproportionate equity splits 
have made themselves visible. From my experience, both equity-split scenarios can lead to 
suboptimal behavior." 

Attracting investors is usually associated with an increased chance of organizational survival and a 
reduced liability of newness.* However, my research with Matt Marx suggests that founding- team 
turnover actually increases dramatically when the startup raises its first round of financing; success at 
atfracting investors is associated with having founding teams become less, rather than more, stable. 
This might be because investors may force renegotiation of equity stakes and roles within the 
founding team, which is almost guaranteed to raise the tensions within the team"^^ This important 
hidden cost of attracting investors should be considered before hunting for capital. 

Costs of Venture Capital: Losing Control of Decision Making 



Although the board can provide guidance to the executive team and monitor the team's progress, its 
most ftindamental role is often decision making. For instance, boards not only review the startup's 
strategy and performance but also help decide if the strategy should change. They not only evaluate 
the CEO's performance but also decide how the CEO should be compensated and whether or not he 
or she should be replaced (see Chapter 10). Boards may also make decisions about financing (e.g., 
whether or not to raise new rounds or to change financing approaches) and about when and how to 
exit. Having a say in all of these decisions is a key way in which investors protect their investments. 
However, all of these decisions create the potential for divergence between the wishes of the 
founders and the wishes of the investors, making the board a key arena in which founders and 
investors wrestle for control. 

Board Control 

As described in Chapter 5, a startup's board tends to start out dominated by the founders. With each 
round of financing, however, the lead investor is likely to want a board seat. Sometimes a new seat is 
added, but more often, to avoid having the board grow unwieldy, the lead investor replaces a founder 
or an early angel investor on the board. In my dataset, after the A-round of financing, founders were 
already a minority within the average board, holding 34% of the seats, while outside directors 
already held 59%. After the B-round, founders were down to 2\% of the seats, with 12% held by 
outside directors, the clear majority of whom were investors.'^'' The board's changing composition 
often becomes a critical issue that affects how much control the founders will continue to have over 
certain decisions. 

Boards usually start out including the founder-CEO along with one or more other founders. A 
founder-CEO who wants to stay on the board can push for his or her seat to be defined as a "founder 
seat," in which he or she will still sit even if replaced by a nonfounding CEO, rather than as a "CEO 
seaf ' that will be taken by the new CEO, possibly leaving the founder without any seat on the board. 
When FeedBurner was raising its A-round, the founding team negotiated that its newly formed three- 
person board would include, in addition to the lead VC, two founders rather than one founder and the 
CEO. Founder-CEO Dick Costolo, knowing that he could be replaced as CEO by a nonfounder, 
pushed for this provision so that he and the team would have more sustained representation on the 
board. 

In some cases, the VCs' presence on the board may lead to much- needed professionalization and 
discipline within the executive team.^^ The VCs may also help shape the startup's strategy in 
productive directions. In other cases, though, it may constrain the founders from adjusting their 
strategy or making other changes. For instance, Evan Williams believed that Apple would beat Odeo 
to market with a superior podcasting technology, so he wanted to abandon the podcasting strategy and 
quickly develop a new status-update tool. Evan's VC-controlled board, however, was starting to 
worry about its investment and about Odeo's lack of progress. "It's the board's job to question what I 
am doing, but I don't want to be required to explain myself [so much]," Evan said. "Having a board is 
killing our ability to try new things." 

Although many board decisions are decided by majority rule, investors may negotiate 
"supermajority" or veto rights that give them extra power when deciding specific issues, such as 
whether to sell the startup. In the same way that liquidation preferences give investors more 
ownership rights than they would have based strictly on the percentage of equity they own (as we saw 



above), supermajority or veto terms can give them greater control rights than would be indicated by 
their percentage of ownership. At one startup, even though the founders retained board control, the 
VCs negotiated veto rights over certain decisions. One of those was approval of the annual budget. 
Years later, during the dot-com crash, the VCs wanted to exit and tried to force the startup to accept 
an acquisition by vetoing all proposed budgets. One founder said, "Their vetoes made it inpossible 
for us to give our enployees pay increases or bonuses and it really was holding us back," forcing the 
conpany to find new investors to buy out the VC firm 

During the downturn of 2008-2009, some founders complained that their boards had become risk- 
averse and opposed to any actions that would require additional resources, preventing the founders 
fi-om taking advantage of the downturn by acquiring competitors who had fallen on hard times and 
fi'om making other aggressive competitive moves. At the other extreme, because VCs try to pick out 
startups with the potential to be very large, they often push those startups to adopt grander — and 
sometimes riskier — strategies. One of the VCs on the board at Lynx Solutions, for example, pushed 
the company to make a radical shift away from its consumer-based product and focus instead on an 
enterprise-based system in order to capitalize on the "hof market for enterprise technology, even 
though its founders believed this would be a major mistake. 

Boards often aspire to make unanimous decisions that have everyone's support. However, for many 
decisions, this is unrealistic and may not even be desirable (if it leads to lowest-common- 
denominator decisions), resulting in board votes that are much more split. Sometimes, every seat 
counts: 

• SpKt board, plus one independent director — Many boards have equal numbers of insiders 
and outsiders, with one additional seat held by an independent director (i.e., a director who is 
not a major investor, current executive, or founder). A three-person board would have one 
founder, one investor, and the independent director; a five-person board would have two 
founders, two investors, and the independent director. On boards with more than three 
members, founders are usually aligned with each other, as are investors, making the 
independent director the "swing vote" whenever a decision is to be made by majority rule. Of 
course, this heightens the stakes in the negotiation over who gets to choose that independent 
director. After FeedBurner's B-round of fiinding, the company established a five-person 
board: two founder seats, two VC seats, and an independent board member to be nominated 
by founder Dick Costolo. Dick nominated Matt Blumberg, the founder-CEO of another high- 
tech startup. "I love having another CEO on the board," explained Dick. "Matt's the only 
person involved in the company who knows exactly what I'm dealing with, is my peer, and 
generally feels the same way I do about the next steps in the life of the conpany." When VCs 
choose the independent director, such alignment between that director and the founders is 
much less likely. 

• Evenly split boards and board gridlock — Many startup boards are evenly split with no 
independent tiebreaker, despite the obvious risk of board gridlock. In fact, after the A-round 
of investment, a fiill 46% of boards have an even number of members (24% have four 
members and 11% have six). This percentage stays relatively constant through the first four 
rounds of financing, ranging from 43% to 49%. Boards with only four members — two insiders 
(founders or early executive hires) and two outsiders (often investors) — are most common 
during the first two rounds of financing (before outside investors gain control of more than 
half the board) and are most prone to gridlock.* 



Managing the Board: The Middle of the "Hourglass " 

The CEO plays a central role at every board meeting. However, a board meeting is not just an event, 
it is a process. The founder-CEO has to prepare for the board meeting, conduct the meeting, and 
perform follow-up tasks after the meeting. All of this comes on top of his or her duties managing the 
executive team. FeedBurner's Dick Costolo described this two-sided challenge: "As founder-CEO, 
you're alone at the center of an hourglass-shaped reporting structure in which all the employees 
report up to you and then you report up to all the investors/shareholders." New founder-CEOs are 
often surprised by the amount of time it takes to perform their "upward" board-management tasks. 
Masergy's Barry Nails recounted the enormous amount of time he spent preparing for board meetings: 
"The preparation of my first board package took a ftill week. We had monthly board meetings, and 
after one of them, I realized I had spent 25% of my month getting ready for it. The amount of work to 
prepare for a board meeting was stunning!" 

Although founders can learn how to perform these tasks more efficiently, the challenge of managing 
the board continues to grow as the board grows. As Jim Triandiflou of Ockham Technologies said 
about the perils of having a large board of directors, "You spend all your time managing your board 

members. "^^ Serial entrepreneur Furqan Nazeeri commented about boards in young startups, "I have 
found that a five-person board for early-stage (A-round) companies is too large. The CEO will spend 
as much time managing the board as he or she will managing the team, which is probably the same 
size. It's also very difficult to remove a board member, particularly a VC, as the conpany grows. It's 
not uncommon for B-round and later- stage VC-backed companies that start with five-member boards 
to grow to six- or even seven-member boards, which is just too cumbersome." 

Lynx Solutions experienced this very problem It was raising a new round of financing and its 
board had grown to four investors and three founders. The founders resisted growing the board to 
accommodate new investors and instead scaled it back to five members. Founder James Malmo said, 
"If we hadn't cut back, we would have been a tiny company with a board fit for GM. ... We already 
had a minority from the management group anyway, and that wouldn't change. . . . [Cutting back to 
five] would give us room to grow. It felt minor league to have just VCs and management on the board; 
a serious VC-backed company brings in prominent outsiders who can add value. That was part of the 

rationale: Leave room to bring on someone from the outside world."^^ 

The Effects of Board Formalization 

When the fiill board is making decisions together, members who are founders or inside executives can 
have an important impact. However, just as startups become more formalized as they grow, boards 
also become more formalized as they grow. In particular, they usually adopt a committee structure, in 
which subsets of directors serve on specialized committees making decisions about specific issues. 
For instance, a conpensation committee may be responsible for setting (or recommending changes to) 
executive compensation and an audit committee may be responsible for monitoring the quality of the 
startup's performance, accounting, and tax data. These committees are often made up of outside 
directors with the appropriate experience or expertise, and some committees (e.g., the conpensation 
committee, which recommends executive compensation) cannot include founders who are still 
executives, potentially reducing the impact of founders and inside executives on those decisions. 
Within my dataset, after the first round of financing, 30% of boards had established formal 



conpensation committees and 21% had established formal audit committees. These percentages rose 
steadily and, by the fifth round of financing, 42% of boards had condensation committees and 34% 
had audit committees. 

Staging as a Control Mechanism 

Gaining control of the board is the most direct way for VCs to gain control over decision making. A 
less direct but still powerfiil way is by "staging" their investments; that is, committing small portions 
of capital sequentially, rather than committing up-front the fiill amount of capital that will be needed 
by the startup. VCs know that they will have the most leverage over a startup when it needs their 
money to grow or just to continue operating. 

Before the initial round of financing, a VC firm's due diligence might reveal risks that it wants 
addressed before it agrees to invest; the VC may therefore require such changes as a condition of its 
initial investment. Each subsequent round of financing also provides the VC with a key lever of 
control over the startup. Before each new round, the VC reassesses the startup's prospects and 
decides whether to reinvest in the next round, to make new demands before reinvesting, or to abandon 
the investment. If the VC decides to reinvest, it will try to provide enough capital for the startup to 
achieve its next major milestone. This "staging" of capital allows the VC to learn more about the 
startup's prospects and to reduce its own risks (and potential agency problems) by not having to 
commit all of the necessary capital up-front.^ ^ 

A key decision for both startups and investors is how long to wait between rounds of financing. 
The technology ventures in my dataset averaged 15 months between rounds; the life sciences ventures 
averaged 17 months.* On the investor's side. Professor Paul Gonpers has shown that VCs are willing 
to allow more time between rounds when the startup has a higher percentage of tangible assets; 
because such startups pose lower agency costs and fewer information problems, VCs can invest for 
longer periods. On the other hand, they allow less time between rounds when the startup is more 
R&D intensive and dependent on intangible assets; such startups pose higher agency costs and more 
information problems. On the startup's side. Professor Bill Sahlman suggests that how long 
entrepreneurs wait between rounds of financing is partly determined by the balance they strike 
between fear and greed; that is, the fear of running out of cash and having to shut down versus the 
desire to keep a higher percentage of the equity by waiting to make more progress and increase the 
startup's valuation. Interestingly, founders suggest that the decision is also partly based on their 
respect for the "disciplining effecf ' of not having a wealth of resources and their fear of the bad 
habits that can develop when a startup has raised too much capital. 

DIVERGENT PATHS: WEALTH VERSUS CONTROL WITH INVESTORS 

Founders face a wide spectrum of investor choices that can pose stark trade-offs. Perhaps even more 
so than for the previous dilemmas examined in this book, each investor choice can enable or imperil 
very different things for the founders. For instance, resisting investments from professional investors 
can enable the founders to keep control of the board and of key decisions, but imperil their ability to 
fill holes in the startup's human capital, social capital, and financial capital, thus making it harder to 
quickly bring a better product or service to market. In contrast, taking money from outside investors 
can enable the founders to tap into the resources needed to grow value, but at the expense of 



imperiling control. 

For instance, when Ockham Technologies was faced with a choice between taking money from 
informal angel investors whose only real contribution would be capital and taking money from a top 
regional venture capitalist, founder Jim Triandiflou had to carefiilly weigh the potential benefits and 
pitfalls of each path. On the one hand, the Texas angel investors required less total equity while 
investing more money ($10 million), promised to stay out of the day-to-day management of the 
conpany, and would accept a small board on which they had only one representative. The drawback 
to the angels was their complete lack of industry and startup knowledge. If Ockham's founders took 
the angel money, they would be completely on their own. 

On the other hand, Noro-Moseley, one of the most reputable VC firms in its region, offered only $2 
million, wanted a liquidation preference, gave Ockham a lower valuation, and required a board of 
directors with five seats, of which only two seats would be held by the founders. (One seat would be 
held by an independent director.) Jim wanted to avoid losing control of the board. But Noro was a 
proven player, had partners with direct experience in Ockham's industry, and would give Ockham the 
credibility that a young startup needed to attract the best employees and to sell its software to 
customers such as IBM. 

Jim, a first-time founder, explained how he and his cofounders made their choice: "We weren't 
overly confident that $10 million from the Texas guys would be enough. If things happened that we 
didn't expect, the Texas guys might run for the hills, but for Noro, this was their business. The terms 
of the Noro deal were the least attractive, but we wanted to get the smartest, best people we could, 
and we didn't let our equity 'delusions' stand in the way of that." In the horse race between fear and 
greed, Jim's fear of running out of cash, given the uncertainties faced by a first-time entrepreneur, 
won out over his "greed" for a higher percentage of equity. 

As we will see in Chapter 1 1, different founders should be making very different investor choices, 
informed by their own core motivations. With each type of investor, founders should also make very 
different choices about the specific investment terms to which they agree. For instance, we focused 
above on how board composition will affect control of the startup, but founders who are particularly 
concerned with keeping control of their startups should also fight back most strenuously against 
protective provisions that give investors extra voting rights and drag-along rights that determine 
who can authorize the sale of the startup to another company. Prioritizing these control-related terms 
will increase the founder's chances of remaining king of the startup. Founders who are less concerned 
with control rights than with maximizing their financial value should be willing to accept the 
provisions listed above and instead fight to maximize their ownership stake overall and in various 
exit scenarios, thus increasing their chances of achieving a Rich outcome. In addition to the 
liquidation preferences described above, they should fight to minimize secondary terms, such as the 
investors' anti-dilution protection (which protects the investors' ownership stake in a subsequent 
down round of financing) and the dividends that investors can accrue. At Ockham, for example, Noro- 
Moseley wanted an "accruing dividend" that would lock in an 18% annual return before the company 
could make payouts to other shareholders.^^ 

CLOSING REMARKS 

Vigilance about investor decisions is particularly important, given (a) the potential disconnects 
between taking outside capital and maintaining control of the startup and (b) the natural inclinations 



that can get founders into trouble when making investor decisions. Prior to "signing on the dotted 
line," founders should understand the repercussions of taking outside money, guard against their own 
optimism, and arm themselves with negotiating leverage whenever possible. 

One of the major repercussions of taking outside money arises when investors move to gain control 
over key decisions in order to protect their investments, imperiling the founder's control over the 
startup. For founders who are wealth-motivated, gaining confidence that a VC will be able to add 
value to the startup often makes it easier to relinquish some control to the VC. But for other founders, 
their control motivation may not be compatible with outside investment and they should be very 
hesitant to raise capital or, if it is absolutely necessary, should try to craft terms that will give them 
the maximum amount of control for as long as possible. ^-^ Investors are usually more attracted to 
startups with wealth-motivated founders, with whom they will be better aligned than with control- 
motivated founders, so founders who truly want — and understand the repercussions of — outside 
ftinding should check that they are sending the appropriate "wealth-motivated" signals to potential 
investors.* 

A founder's natural optimism can also inttoduce long-term problems when taking early investor 
money. Founders who agree to standard investment terms — for instance, to give their investors a 
liquidation preference — ^because they are confident that their startups will grow big, are giving up 
claim to a moderate gain. Yet it is much more likely that a startup will grow to moderate size than to 
huge size. The Lynx Solutions team, for example, agreed to a liquidation preference for their A-round 
that led to problems three rounds later when they were weighing acquisition offers that were below 
their cumulative liquidation preferences and thus would not give the founders any financial return 
from their years of hard work. Instead of assuming that such a term will never apply because the 
startup is sure to become big, optimistic founders should take the small-exit scenario seriously from 
the beginning and should consider whether they really are willing to give up any possible gains from 

such an exit."'" 

Dealing with investors can be daunting, particularly for new enttepreneurs, but there are several 
sttategies that founders can use to gain leverage. Jim Triandiflou deliberately waited until his startup 
had a solid product, cash flow, and customer conttacts before approaching VCs, judging correctly that 
this strategy would result in higher valuations. He also avoided working with a single VC firm, in 
order to avoid the conttol over decision making that a single VC with two board seats would have. 
Instead, he split the first round between two VC firms, each of which took one seat on the board. 
"Two guys from the same firm can monopolize decisions, and not giving any one person or group 
conttol is critical," Jim explained. "My having board conttol is less important than my making sure 
someone else doesn't have it." 

For FeedBurner's Dick Costolo, gaining leverage in a VC negotiation meant having appealing 
alternatives. In one round, he had gained leverage by getting competing term sheets from different 
VCs. In another round, he put off approaching VCs until he had received acquisition offers from 
Yahoo and Google: "Having the potential acquirers involved changed the dynamic with potential 
investors. It kept acquirers from wasting our time with stupid offers and prevented a lot of back and 
forth with the VCs on harsh terms." Dick also made sure that he negotiated with VCs well in advance 
of running out of cash: "If we take it down to the wire, the VC will sit on it for a few weeks until they 
know we don't have money, have to sign their term sheet, and have to take their terms." 

As we have seen, core founders are often missing skills, connections, and financial resources 
needed to build the most valuable startup possible. They can make up for this by atttacting 



conplementary cofounders or by hiring talented nonfounders. A third possibility is to add investors to 
the team, but founders must first think carefiilly about the often-hidden implications for themselves, 
for the startup, and for the board of directors. For a founder-CEO, losing control of the board is only 
the first step on the way to losing control of the major decisions in the startup and may culminate in 
one of the most critical inflection points in the life of a startup, the replacement of the founder-CEO 
by a nonfounding CEO. That inflection point is the subject of our next chapter. 



* A liquidation preference is a term agreed to in the initial financing deal between venture capitalists and entrepreneurs that governs 
the division of profits in the event of the venture's acquisition or liquidation. We delve into liquidation preferences later in this chapter. 

* Likewise, a distinct group of technology founder-CEOs, such as Joel Spolsky of Fog Creek Software, proclaim their aversion to 
taking money from VCs and similar investors (e.g., Spolsky, 2003). 

* At the same time, attracting one type of early investor may preclude the attraction of a different type later. For instance, taking 
money from — and/or giving board seats to — some family members or angel investors may turn off a VC or lead to tough negotiations 
over the VCs desire for those relatives or angels to leave the board, something they might be loath to do. 

* The connections and credibility added by top-tier angel investors can exceed those of many VCs. For instance, since 2007, angel 
investor Ron Conway, one of the first to invest in Google, PayPal, and many other successful startups, has been among the top 10 
investors according to the Forbes "Midas" rankings, higher-ranked than the vast majority of VCs. 

* It should be noted that, due to problems with the transparency of the angel market, the angel data most probably understate the 
actual number of angel investments (Aldrich, 2010). 

* As shown in Figure 9.4, 21 % of the startups in my dataset initially approached their VCs by submitting a business plan. In 18% of 
initial VC rounds, the founder had already worked with the VC in a prior venture. 

* Boards also often include "observers" — people who can come to board meetings but usually do not actively participate in them and 
do not vote on decisions. As a condition of their investments, VCs and other investors may negotiate the right to have an observer on the 
board, but those observers do not have a say in board decisions. 

* Future research could shed light on the VC side of this, examining whether the different backgrounds of VCs result in different 
levels of engagement with their startups, different types of value addition, and other factors. 

* The results described here regarding formal board meetings were statistically significant even after including a proxy for the 
frequency of the informal board-founder interactions. 

* Further analysis of down rounds shows that 4% of B-rounds were down rounds (Le., at a lower valuation than the A-round) and 7% 
of C-rounds had a lower valuation than the B-rounds, followed by 9% for the D-rounds and E-rounds. 

* A related term, which can also affect the founders' eventual payouts, is whether the investors' preference includes participation. 
After the initial preference is paid out to investors, participation governs how the remaining proceeds are paid out. There can be full 
participation (investors share pro rata in the remaining proceeds as if they had converted their preferred shares into common stock), no 
participation (investors have to choose whether to keep their liquidation preference or to convert their preferred stock to common stock), 
or capped participation (a midrange option). For more details on this term and for the potential misalignments that each option can cause, 
see Wilmerding (2004) and Bussgang (2010). 

* The impasse was resolved only when the investors finally agreed to create a small "equity carve-ouf that would give the founders 
and some key personnel small shares in the sale. 

* Among VC-backed startups, 75% of founders received no financial return from their years of hard work building the startup (Hall et 
al., 2010). 

* For more on the liability of newness and Stinchcombe's (1965) arguments about the factors that underlie it, see the beginning of 
Chapter 1. 

* To avoid such decision-making gridlock, the millennia-old Talmud goes out of its way to ensure that no court ever has an even 
number of judges. At the beginning of its main volume on jurisprudence (TB Sanhedrin, p. 2a), the Talmud outlines nearly three dozen 
widely varying types of court cases, from small monetary claims to capital cases, which are judged by courts ranging in size Irom 3 to 71 
judges. The Talmud mandates that every single one of those courts have an odd number of judges. 

* The time between rounds was very consistent across levels of startup maturity: In technology startups, the intervals were 15 months 
between A-round and B-round, 16 months between B-round and C-round, 16 months between C-round and D-round, and 14 months 
between D-round and E-round. Within life sciences startups, the intervals were 17 months, 17 months, 15 months, and 16 months, 
respectively. 

* Term sheets and investment deals should be formulated different^ for wealth-motivated founders (who, for example, are likely to be 
more open to being replaced by a better CEO) and for control-motivated founders (who, for example, may be averse to accepting the 



necessary outside capital unless their control of the startup can be solidified). Adeo Ressi's influential founder-oriented website, 
TheFunded.com, has created a set of "founder friendly" model term sheets to complement model term sheets created by the National 
Venture Capital Association. Such models make it possible for founders and investors to create a coherent spectrum of choices that 
recognizes how founders differ regarding control-versus-wealth trade-offs and how those different incentives can be aligned with those 
of investors. 

^ Investors could help founders avoid some of the ill-fated decisions that result when founders default to natural human tendencies 
rather than consciously deciding what to do. For instance, as examined in Chapter 4, founders commonly found with friends or family — a 
very natural tendency — but fail to erect firewalls to make sure that the team does not blow up (and take the founder's most treasured 
relationships with it). Likewise, as examined in Chapter 6, natural human inclinations lead founders to avoid structuring dynamic equity 
agreements, despite the probability that static agreements will harm the team and the startup. Investors could offer creative incentives to 
avoid such missteps; for example, by making the implementation of specific firewalls or dynamic agreements a condition of investment. 



CHAPTER TEN 

FAILURE, SUCCESS, AND FOUNDER-CEO 
SUCCESSION 



Thus far, we have examined founders making their very first decisions, from becoming founders 
early or late in their careers, to going solo versus building a founding team, to making choices about 
hiring and about attracting investors. Through the experiences of Evan Williams and a host of other 
entrepreneurs, we have become acquainted with the many ways founders can make mistakes. In this 
chapter, we turn to one of the most inportant outcomes affected by early decisions, which also 
happens to be one of the most critical points in the life of a startup and of its founder — the point at 
which the founder-CEO is replaced as CEO. As we will see, although some founders, such as Evan 
Williams, are able to orchestrate their own successions, for many others, it comes as a shock. 

In contrast to some of the founders we have seen. Lew Cime avoided many early pitfalls. From the 
very beginning of his career. Lew made choices that would enable him to start and grow a successful 
startup. After earning a degree in conputer science at Dartmouth, he furthered his technical education 
by taking a job at Apple Computer. While performing his core tasks there, he also conceived an idea 
for a diagnostic tool for Java-based enterprise systems and began to make plans to launch his own 
high-tech startup. He decided to gain more experience in both technology and management while also 
getting an insider's view of a startup by working for two years at Hummingbird, a young IT conpany. 
"I wanted to learn how to be a founder or an early enployee of a new conpany," said Lew. "I wanted 
to gain breadth, conpared to my more focused role at Apple. . . . Part of my motivation to start [a 
company] was to grow professionally by achieving as a businessperson, not as a technologist." While 
at Hummingbird, Lew picked the brain of a former entrepreneur who helped him hone his idea and 
gave him insights on hiring and fund-raising. 

Lew decided to fly solo, founding Wily Technology on his own and raising $100,000 from angel 
investors. From the beginning, he was passionately attached to his startup and made personal 
sacrifices designed to benefit Wily. Lew worked alone and feverishly for a full year to develop his 
Java-based technology. He sent a white paper about his idea to IBM, which resulted in Wily's first 
sale. Lew then hired a "generalist" who could take on the nontechnical tasks while Lew himself 
continued to work on the product. When his early board members (the angel investors) dragged their 
feet on giving the new employee equity, Lew carved out some of his own equity. "I transferred [some 
of my shares] to [my enployee]," he explained, "rather than dilute the rest of the shareholders." The 
principle Lew was following was, "Don't worry about your own dilution, worry about Wily." 

As Lew continued to refine the product and the business plan and bought his first suit so he could 
make sales calls, he realized he needed to raise more money for fiill-scale product development. 
Based on the strength of his technology and his vision for the product, he raised $2 million from 
outside investors. He then hired one of his former Apple colleagues to take on the chief scientist role. 
This was "a pretty challenging thing for a technical founder to do," Lew said, but the new version of 



the product benefited from the new leadership. Lew hired more key people, paying them more than he 
himself made as CEO. 

Lew continued to build Wily with help and guidance from his VC investors and the startup 
continued to meet or exceed all its early milestones. Lew was particularly excited as the company 
began shipping its new version of the product and started negotiations on a new round of financing. 
His excitement quickly turned to shock when his backers insisted that he step down as CEO as a 
condition of the new round. "Things became really tough after that," Lew recalled. "All I could think 
was, where have I messed up? What have I been doing so wrong?" In the end. Lew's potential 
replacement as CEO, Richard Williams, refiised to take the job unless Lew also gave up his position 
as chairman of the board. At this point. Lew faced a major dilemma: Refiise to hire the professional 
who could build Wily's value, or hire the pro and be completely sidelined in the startup he had 
nurtured as if it were his only child. 




Founding A-round B-round C-round D-round 



□ On 3rd (or more) CEO 0% 6% 996 17% 23% 

□ On 2nd CEO 0% 19% 29% 35% 38% 
■ Founder still CEO 100% 75% 62% 48% 39% 



Figure 10. 1. Percentage of Startups Still Being Led by the Founder-CEO 



How could this happen to such a successfiil entrepreneur? The move to replace Lew came as Wily 
was raising its third round of financing; we can see in Figure 10. 1 that, by that time, more than half of 
startups have replaced their founder-CEOs. In fact, 17% of them have already replaced CEOs more 
than once. But why? How often is it by choice and how often is it forced on the founder-CEO? What 
happens to the replaced founder? These and related questions are central to one of the most critical 
inflection points in the evolution of a startup: the "succession" from a founder-CEO to a 
"professional" (nonfounding) CEO. 



Founding 



Decision 
to Replace 



New CEO 
Enters 



Y 

Antecedents 

• Degree of success 

• Rate of growth/change 

• Loss of board control 

• Founder capabilities and centrality 

• Founder burnout/motivation 

• Trigger of succession: founder, board, others 



Y 

Search 

• Sources of successors 

• Founder involvement 

• Successor dissimilarity 



r 

Post-succession 

• Retain replaced founder? 

• Lower-level executive? 

• Remain on board (chairman?) 

• Successor taking control 

• Gaining predecessor's support 



Figure 10.2. The Founder-CEO Succession Process 



Throughout this book, we have focused on startups in which a founder is still CEO. However, a 
startup's first-ever CEO transition is a particularly challenging and jarring juncture. To help us delve 
into founder-CEO succession, we will both examine large-scale data about the antecedents, 
dynamics, and aftermath of such events and delve into the stories of several founder-CEOs and the 
professional CEOs who replaced them. We will analyze why successfiil founder-CEOs like Lew get 
fired as CEO, observe in detail the process of replacing them, and explore whether they should 
remain active in the startup. Figure 10.2 is a graphical depiction of the most important stages of the 
founder-CEO succession process and some of the factors affecting each stage. 

Although we have already seen many decisions in which control-motivated founders diverge from 
wealth-motivated founders, the divergence is most stark during the founder-CEO succession process, 
affecting who triggers the succession, how the search for a new CEO is conducted, and what — if any 
— role the replaced founder-CEO will play in the startup. 

TRIGGERS FOR CHANGE 

Founder-CEO succession can be initiated by the founder-CEO, the board of directors (which is 
officially responsible for selecting and monitoring the CEO), or some other party, such as investors 
who do not serve on the board. In my dataset, founder-initiated changes were relatively uncommon, 
accounting for only 27% of succession events.^ The lion's share of CEO changes — 73% — were 
initiated by the board or by another party.* Board-triggered successions are usually the most tense. 
The founder-CEO often resists the change, heightening the risk to the startup. After examining the 
smoothest type of succession — voluntary succession triggered by the founder — ^we will look at two 
types of board- initiated succession: first, when the board fires a poorly performing founder-CEO and, 
second, when the board replaces a high-performing founder-CEO. The latter confronts us with what is 
known as the paradox of entrepreneurial success. 

Voluntary Succession 

\bluntary succession is when the founder-CEO decides to step aside for a new CEO, who almost 
always comes from outside the startup. This is usually the least stressfiil transition, but given the tight 
attachment between founders and their startups, and the typically high level of confidence founders 
have in their skills, it is a rare founder who willingly gives up leadership of his or her startup. As 
Lew Cirne said, "Prior to Richard [the professional CEO] coming along, I felt it was my company 



— 'Lew's company.' " For such a founder, voluntarily handing over the reins of the startup is as 
inconceivable as voluntarily putting up for adoption the child he or she has raised since birth. 

Founders who willingly volunteer to be replaced may be those who are most self-aware and have 
become convinced that the demands of the CEO job are beyond their abilities.* Alternatively, they 
may be those who feel burned out from riding the entrepreneurial roller coaster, putting in seven-day 
weeks month after month. In either case, the founder begins to realize that the value of the startup is 
going to suffer. By the time Mike Brody founder-CEO of Transcentive, told his replacement, Les 
Trachtman, that he was in favor of stepping down, he had been CEO for over fourteen years and was, 
as Les explained, "a little older, a little more tired," and motivated to hand Transcentive over to a 
professional CEO who could lead it more effectively. 

Founders primarily motivated by building wealth are often quicker to step aside as CEO when they 
see the startup's value — and therefore the value of their equity — suffering. In contrast, control- 
motivated founders will be more likely to fight for their CEO positions and thus end up having the 
board initiate the change. As we will see below, waiting for the board to trigger the change can have 
important consequences for the founder's role after succession. Founders who maintain some control 
over the succession event often end up in better positions. For example, Mike Brody at Transcentive 
was able to retain his position as chairman of the board, largely because he took an active role in his 
own succession. 

Fired for Failure 

A board may trigger CEO succession if the startup has been performing poorly under the founder- 
CEO's leadership and the board believes a new CEO could do better.^ (Such a situation is typical in 
large companies, where firing the CEO almost always comes after underperformance or a failure to 
meet expectations.) VC Tim Connors says that, in startups, boards "look for a CEO to raise money, 
hire a team, and set and hit a financial plan," and that failure in those areas will cause them to replace 
the CEO. 

In startups, however, failure is harder to judge than in big conpanies, especially when the startup 
has not yet completed product development. There are no sales or revenue figures, customer 
acquisition has not begun, and other market-driven metrics are not available. However, startups do 
have a variety of milestones — such as completing development of each major module of the system or 
hiring an experienced VP of marketing — that they are actively tracking and sharing with the board. 
When those milestones are missed, or missed one after the other, the board may conclude that the 
leadership team, particularly the CEO, is dropping the ball, whether by choosing the wrong 
milestones, inaccurately forecasting when they will be achieved, or choosing the wrong processes or 
people to achieve them. One founder-CEO, for instance, came up with an industry- changing idea, 
launched his startup, developed a prototype, and raised $2 million in two rounds of financing. 
Nevertheless, he was fired by his investors after more than a year of consistently missing product- 
launch targets and was replaced with a successor who got the company back on target. 

The Paradox of Entrepreneurial Success 

Lew Cirne was neither a volunteer nor a failure. He was enjoying his role as CEO, he felt confident 
he was up to the job, and Wily was unquestionably doing well. By almost any metric. Lew was a 



smashing success as CEO. He had built a 50-person company and led it through a very successful 
product-development process. As a result, Wily's revenues were rising quickly and the company was 
headed toward profitability. To get the "rocket fiiel" to invest in product development and hire his 
team, he had raised two large rounds of financing from venture capitalists — and top-tier ones at that. 
Attracting rocket fiiel was not only an essential ingredient of Wily's success, but also an important 
certification of that success; VC investments serve as both a stamp of approval for the startup's 
progress and a boost to its credibility.-^ One investor, Greylock Venture Partners, went so far as to 
state on its website its "belief in the entrepreneur as 'Star' and Greylock as 'Invited Guest,' " and 
Lew certainly had every right to feel like a star entrepreneur. Like many founders, he referred to his 
startup as his "baby." He was brimming with pride in its growth and enjoyed being at the center of its 
operations. 

Why, then, was the father of this baby being fired by his "invited guesf ? 

The seeds of such founder-CEO succession events are planted much earlier. Quantitative analyses 
of my data along with my field research highlight two types of success — product development and 
fiind-raising — that spark crucial but underappreciated internal changes that, in turn, put successfiil 
founders such as Lew at risk of being fired.^ 

Succeeding at Product Development 

The major operating challenge in the initial phase of a startup is to develop a product or service that 
can be sold to customers. This requires the accomplishment of many complex steps: defining 
specifications, developing an architecture, surmounting technical or scientific hurdles in 
development, coordinating team efforts to develop different yet intersecting pieces of the product or 
service, gaining any necessary regulatory approvals, and the like. At this stage, there is a premium on 
technical or scientific skills and on leadership at the project- team level. For Wily, Lew Cirne's 
expertise was particularly important in the fast-developing world of Java applications. When the 
original client-side product didn't get the traction Lew had anticipated, he was able to quickly 
regroup and redesign the product in time to catch the growing server- side wave. Founders with 
relevant technical or scientific backgrounds, such as Lew, are often the best leaders for the product- 
development phase of the startup, for they are best suited to address the technical contingencies and 
challenges of that stage. 

However, once the product or service has been developed and is ready to be sold, the startup 
becomes much more complex and the CEO faces dramatically different challenges. The startup is no 
longer primarily a technical team developing a product. It must add new tunctions such sales, 
marketing, and customer support. A technical founder-CEO such as Lew, who had to buy his first suit 
in order to make a sales call, now has to interview and hire salespeople, understand how their 
motivations differ from those of technical whizzes, and know how to structure their incentives and 
conpensation. Few founders who were adept at the early technical challenges are equally — or even 
sufficientiy — adept at these very different challenges. Managing a technical team is quite different 
from managing multiple fiinctions that must interact and with most of which the CEO has little direct 
experience. At this point, the startup's finances and metrics also become much more complex, 
requiring a level of financial sophistication possessed by few founder- CEOs.* 

The leap from leading product development to leading a multifiinction startup challenges not only 
the founder's skills, but — perhaps even more profoundly — ^his or her values. During the early days. 



when founder-CEOs are still relying heavily on their own networks for hiring (see Chapter 8), they 
tend to bring in people they already know well and build a tight-knit culture. "Whom we hired was 
important," Lew insisted, "because we were trying to build a culture that felt like a family." Founder- 
CEOs therefore tend to exhibit fierce loyalty to their early employees. Those early employees often 
prove unable to meet the challenges of the startup's next stage of growth, yet the founder-CEO, prizing 
loyalty and not wanting to imperil the carefiilly constructed family culture, often keeps them in key 
positions long after they have reached the limits of their capabilities. The startup has reached the 
point at which those early employees must be managed much more objectively than the founder-CEO 
can do. After professional CEO Les Trachtman succeeded founder-CEO Mike Brody at Transcentive, 
he reflected on how hard it would have been for Mike to make the personnel changes that the startup 
required at that stage of growth, particularly to replace his own brother, who had long ago ceased to 
be an effective CFO. "[Mike was in] a tough situation," Les recalled. "A couple of people he knew 
were in the wrong positions, but because he valued his personal relationships with those people, he 
didn't want to make the moves he knew were necessary. I think I was brought into the firm because 
the board realized that Mike couldn't jettison the staff that needed to be let go. He had developed such 
strong ties with his employees. ... I [also] think Mike knew that his brother was not the right person 
for the CFO job," but couldn't bring himself to change CFOs. 

The founder-CEO may also be attached to his or her original idea for the startup, while the startup 
itself has reached the point at which the original idea may have to be adjusted. At Megaserver, an 
earlier startup at which Les Trachtman had worked, the technologist-founders had spent years 
developing their idea for a "virtual superconputer" and kept finding ways to give it more 
fimctionality and add "bells and whistles." Finally, the startup's chairman decided that it was time to 
move past the stage of developing "cool technology" and to get to work developing a usable system 
that customers could understand and would want to buy, a change in emphasis that the founders would 
have trouble accepting. To help effect such a change, the chairman hired Les, a sales-oriented 
technology executive, to be Megaserver 's first nonfounding CEO. As Les explained, "We needed to 
highlight a specific fimctionality of the technology so that we would not confiise consumers. The 
[founders], on the other hand, felt that by taking this route, consumers would never know that the 
technology had a lot of other great fimctionality." 

The next stage of the startup's development often requires the development of new processes, the 
formalization of the organizational structure, and other changes foreign — or even antithetical — to the 
typical founder-CEO. One influential model of startup evolution emphasizes that each stage of startup 
development has a very different "dominant problem" that the startup must solve, but that the solution 

often causes the startup's next crisis.^ For instance, solving the early "creativity" challenge in a 
startup leads to a crisis of leadership, which is solved in the next stage by an increase in direction, 
which in turn causes a crisis of autonomy. The skills and personal characteristics needed to solve one 
stage's dominant problem often become obstacles to solving the next stage's dominant problem. The 
radical differences between stages require radical shifts in the capabilities of the executive team* 

To add to the injustice, the more quickly a founder-CEO leads the startup through these early 
stages, the more quickly he or she will need to develop new skills or capabilities and the sooner he 
or she may find the challenges winning the race. Thus, the founder-CEO's very success at leading a 
fast-growing startup through product development is likely to accelerate his or her own obsolescence 
and replacement. Quantitative data show that it is precisely at the point where the startup has 
completed product development that the founder-CEO's chances of being replaced rise significantly. 



Succeeding at Fund-Raising 

For many startups, a key challenge is attracting outside capital. Such capital not only helps the startup 
bring its product or service to market, but — if it comes from professional investors — also increases 
the startup's credibility with potential customers and other external parties."^ Yet meeting the fund- 
raising challenge can set off an even bigger change within the startup. With each round of financing, 
the startup sells equity to outside investors and adjusts the composition of its board accordingly. As 
described in Chapter 9, the founders, once they begin raising capital from outside investors, quickly 
lose control of board-level decisionmaking. Thus, the startup's success at raising outside capital has 
not only accelerated its growth rate, but has also caused a fundamental shift in the power structure 
within the board — ^which hires and fires the CEO. The new board, made up of investors and 
independent directors who are motivated by the creation of shareholder value, may have a very 
different view than the CEO does of who is the best person to tackle the startup's next set of 
challenges. Investors may well believe that a new CEO is needed for the startup to reach its potential 
value during the next stage of growth, and they now control the body that can make that decision. 
Thus, with successful fund-raising comes loss of control of the board and, with it, the risk that the 
founder-CEO will lose his or her own position as CEO. 

For Lew Cirne, the VCs' demand for a new CEO coincided with the raising of Wily's third round 
of financing. Lew's seed round had been with family and friends — angel investors who did little more 
than act as an advisory board. After the first round of VC financing, Wily's board consisted of Lew 
(chairman), an angel investor who had been Lew's mentor and fully supported him, and the VC who 
had led the first round. After the second round of financing, Wily's board gained another VC, the lead 
investor in the new round. It was this newly constituted board that soon insisted that Lew step down 
as CEO. Henry McCance, chairman of Greylock Venture Partners, explained how changes in the 
challenges facing Wily sparked the investors' desire for a new CEO: "When we get into the go-to- 
market phase, when sales-and-marketing becomes a bigger issue, in most cases the entrepreneur 
doesn't have that in his background, as was the case with Lew. So we put it out there that there will 
come a time when we need a CEO with different skill sets." 

In general, the more critical a resource is to a startup, the more the startup must give up to attract 
it.^ If that resource is capital, the startup's founders typically must sacrifice control of the board and 
the founder-CEO must sacrifice control of his or her destiny. Quantitative analyses confirm that the 
chances of founder-CEO succession rise with each new round of financing, the more so the more 
capital is raised in that round. 

Implications for Succession 

As the founders and employees of startups celebrate each milestone, feeling that much nearer to the 
success they envision, they seldom realize that they may also be that much nearer to the day when the 
"fearless leader" who hired them and has brought them this far will be replaced. 

Some founder-CEOs are aware of this possibility — perhaps having already been replaced in a 
previous startup, or having heard stories from their mentors who were — which can introduce 
perverse incentives. As one commented, "If the company tanks, I'm gone. But if it's a big hit, I'm also 
gone. If I want to remain CEO, should I only aim for middling success?" Confirming this founder's 
intuition, young companies are indeed more likely to experience the departure of the founder-CEO 



when their revenue and employment growth are very low or very high; middling growth rates lead to 

the lowest rate of founder-CEO departure.^ Control-oriented founders in particular may prefer to 
sacrifice fast growth in order to grow at a rate closer to their own ability to learn and to adjust to the 
shifting challenges they know they will face. At Sittercity, for example, COO Dan Ratner 
acknowledged that founder-CEO Genevieve Thiers was able to remain CEO for several years 
because of the slow growth Sittercity experienced after being founded during the dot-com bust and 
before it faced real competition. "You don't normally have a four-year run-up to the starting line like 
[Genevieve] has had," explained Dan. "She . . . had the time to learn how to get through 
roadblocks. ... No matter how smart you are, you need a certain amount of time, not just time to 
absorb but to make mistakes." 

Understandably, even wealth-motivated founder-CEOs often disagree with VCs over the right time 
to change CEOs, with VCs preferring an earlier change than do the founder-CEOs. VCs have voiced 
the opinion that, when it comes to changing CEOs, "if in doubt, better to pull the trigger too soon than 
too late," while founder-CEOs almost always prefer to receive the benefit of the doubt as long as 
possible. As a result, CEO changes often occur later than investors prefer, but earlier than founder- 
CEOs prefer. In my data-set, for exanple, VC-triggered successions occurred an average of 6 to 12 
months earlier in the life of the startup (3 to 3.5 years into the startup's life) than did founder- 
triggered successions (approximately 4 years after founding). 

The final challenge introduced by the founder-CEO's success is that this very success makes it 
harder for the founder to believe that a different CEO is needed. After all, the founder's worthiness to 
lead the startup, while it may have been a gamble at the beginning, has now been proven. (In contrast, 
a founder who is failing, drowning in problems he or she doesn't know how to solve and wishing for 
assistance, is usually much more receptive to the earthshaking message from the board of "You're not 
the right person to lead us into the next stage.") Indeed, a very experienced VC told me, "One of the 
toughest jobs I have is firing CEOs who have succeeded in rapidly growing their companies." 

SEARCHING FOR A SUCCESSOR 

About one-third of replaced founder-CEOs leave the executive team within a month after the decision 
has been made to replace them The percentage varies according to which party triggered the change: 
When the board initiates the change, the founder leaves the executive team 37% of the time; when the 
founder-CEO initiates the change, he or she leaves the team 24% of the time. In either case, the 
founder-CEO abruptiy goes from being his baby's very attached parent to having no custody at all. 
But for everyone else, and for the two-thirds of replaced founder-CEOs who remain as non-CEO 
executives, the next step is the search for a new CEO. 

Sources of Successors 

In large companies, a high percentage of new CEOs come from inside the organization. For instance, 
in a broad study of 1,035 large-company succession events, 81%) involved inside successors.^ 
Promoting an inside executive is seen as healthy for the organization and gives the board confidence 
that the new CEO will fit with the organization and its culture. One notable exception is when the 
organization has been performing badly and requires dramatic changes; in such cases, an outside hire 
is much more common.^ 



In startups, however, the new CEO almost always comes from outside the existing executive team 
Li dataset, only a handful of startups chose a successor from within the executive team The new 
CEO is usually being brought in to make major changes, whether in the organizational culture (e.g., 
moving from Lew Cime's "family culture" to a professionalized one), in the strategy (e.g., altering the 
founder's original business idea), or in the team (e.g., replacing cofounders or early hires who aren't 
performing well). In short, the new CEO is being hired to do what the founder-CEO either could not 
do (for lack of skills or knowledge or because of deeply ingrained mental models or schemata) or 
would not do (given his or her nonwealth motivations or his or her attachment to longtime 
participants or to the original strategy or idea). The startup's small executive team seldom offers the 
board enough good candidates. In addition, as noted by venture capitalist Jeff" Bussgang, "Great 
general-management and operational executives don't typically take startup roles running a single 
frinction," requiring the startup to look outside for such executives. 

Wily's search for a new CEO touched on a wide variety of both internal and external sources. Lew 
Cirne's initial inclination was to promote Vic Nyman, the VP of sales and marketing. "Vic was an 
absolutely key person in the growth of Wily. He was a great team player who would take on whatever 
job we needed him to do," said Lew. "At the same time, he was clearly aspiring to be a president or 
CEO. He served the company and the team very well, but all the while accomplishing his own 
objective of gaining breadth in his own career." However, David Strohm, the VC from Greylock who 
served on the Wily board, immediately rejected the idea because he was "not comfortable with the 
internal candidate." David added, "The leadership team at Wily largely consisted of people who had 
grown into jobs within Wily, some without extensive prior experience at leading teams or managing 
fimctions. . . . [Also, Vic's] sales and marketing fimctions were where most of the growth-induced 
conflict was occurring." 

Wily therefore began a search for outside candidates. The startup spent three months engaged with 
a small boutique search firm, but the few candidates it sent were, according to Lew, "disastrous," "a 
bad fit with our culture," and "people who were out of work and should continue to be." Wily then 
searched through its own network of contacts, but "the first candidates we approached didn't want to 
take our call. It was hard to get a high-quality person's attention to even hear our elevator pitch," 
lamented Lew. Several more months went by and Wily engaged another search firm Within a few 
weeks, the new firm had found "solid candidates" for the board and Lew to consider. But after 
looking at 120 resumes and interviewing 20 candidates, the new investor in the upcoming C-round 
instead recommended one of its own "favorite CEOs," Richard ^^^lliams, who had a "reputation as a 
superstar." Thirteen months after the search process started, the board and Lew agreed that Richard 
would be the right choice for CEO. 

^^^ly ended up finding a new, outside candidate, but in some cases, startups have prior familiarity 
with their potential new CEOs. Some hire new CEOs who have served on the startup's board or 
served the startup as a consultant. Transcentive's Les Trachtman, for exanple, had forged a strong 
working relationship with founder-CEO Mike Brody, first as a consultant and then as VP of 
operations, before succeeding Mike when Transcentive's board felt the need for a professional CEO. 
Similarly, Lynx found its professional CEO, Clark Evans, when he came to conduct an all-day 
strategy session. He was well received by the employees and, as a veteran of a large consumer- 
products company, had the business-process skills that Lynx's founders lacked. As founder James 
Milmo recalled, "People breathed a sigh of relief that we founders were smart enough to bring in 
someone like that." 



Founder Involvement 

While the interval between deciding to replace the founder-CEO and hiring a new CEO can be as 
little as a month, this is uncommon. In my dataset of founder-CEO successions, 6% of the startups 
hired a new CEO in the same month that the decision was made to replace the founder-CEO. Much 
more often, the search is drawn out. In fewer than one-third of the startups was the CEO hired within 
a year of the initial decision; Wily's 13 -month search was fairly normal. This is especially true if the 
founder-CEO who is being replaced is playing a central part in the search process. For the outgoing 
founder-CEO, hiring a replacement is sometimes akin to approving a husband for his or her only 
daughter; no one is good enough. 

Legally and formally, it is the board's responsibility to choose the new CEO. But boards often 
make the founder a central part of the search process, in an attempt both to get input about candidates 
from the person who best knows the startup and to help gain the support of the replaced CEO. 
However, when the founder's motivations (e.g., to maintain the culture and ensure that loyal early 
employees are not laid off) conflict with those of the investors (generally, to find a CEO who will 
maximize the financial return from the startup), involving the founder in the search process can 
introduce new risks and delays. "I didn't want us to rush," Lew asserted. "When you have high 
pressure to find the right CEO, you're going to rush into taking the person who 'looks good' on paper 
but doesn't fit too well with your organization and its culture. An important and challenging thing is to 
provide pushback to investors on the fit piece, because they'll be pushing hard on the people whose 
resumes look really good. But the investors aren't the ones who have to spend every day with them — 
we are!" Boards must carefiilly consider what they might gain from a certain level of involvement on 
the outgoing founder-CEO's part versus what that level of involvement might cost. Who will conduct 
the first screening interview of prospective candidates, the founder-CEO or the board? Will the board 
solicit input from the founder-CEO but retain ftill control over the final decision? Will the founder- 
CEO have veto power? 

Lew's intimate involvement with the search process and his intangible criteria for finding the right 
person caused him to reject several candidates favored by the investors, particularly at the beginning 
of the process. Lew e?q)lained, "I believe it's ... the intangibles that will make or break this turning 
point for the company. I was very happy with the people I had hired and the culture we were building 
and wanted someone who wouldn't feel that he needed to put his mark on the conpany, to assert 
control and feel like he had to make changes in the team" Lew also didn't want someone 
overconfident who "hadn't made any mistakes yet . . . so no humility." He vetoed one person who he 
felt "used motivation by fear," and he avoided people who were "younger, very ambitious with the 
constant need to prove [themselves]." 

When the board involves the founder in the search process, CEO candidates have to meet the 
criteria of two very different parties. The candidates have to be aligned with the board members, in 
particular the investors, who are wealth-motivated and are changing CEOs in order to increase the 
value of the startup. The candidates must also be acceptable to the founder-CEO, who may well be 
control-motivated and seeking every way to retain confrol of key decisions. Even if the founder-CEO 
is wealth-motivated, he or she may have other and conflicting motivations, such as emotional 
attachment to the startup or worries that his or her contributions will be forgotten and his or her 
legacy weakened. 

At Wily, the family culture led to even broader involvement in the hiring process by people who 
did not serve on the board. At the beginning of the search process. Lew took an open approach to 



hiring. He invited most of the company's senior managers to sit in on interviews and to read resumes 
to ensure that the team would feel comfortable with the hiring process and that the new CEO would fit 
Wily's culture. But this approach scared off the highest-quality candidates. As David Strohm, Wily's 
VC, explained, "When you include everyone in it, the process quickly becomes a plebiscite, which 
can be quite off-putting to serious candidates. It can also result in a 'least common denominator' 
outcome, with the candidate who is least threatening or most agreeable to everyone on the team 
surviving." 

Successor Dissimilarity 

During the search process, homophilic tendencies — the tendencies for "birds of a feather to flock 
together" — can lead founder-CEOs to feel drawn to candidates who share their backgrounds and 
skills.* At the same time, for many founders, the more different the candidate is, the more acceptable, 
if those differences address important contingencies the founder couldn't address. Thus, a founder- 
CEO who lacks sales experience but whose startup is ready to sell its product will more clearly see 
the value of a candidate with a strong sales background. A founder with 10 years of work experience 
will more clearly see that a candidate with 30 years of experience may be a better choice than a 
candidate with the same 10 years. For Lew, the fact that Richard had a different skill set made the 
change more palatable for him. "What got me comfortable with being on board with [the succession] 
was recognizing that [Richard] couldn't have done the job I had done to get the company to where it 
was. I felt better about myself, that I had done something uniquely valuable for the company. I had 
gotten the company fi*om A to B, but now we have to go from B to Z." 

It is also easier for the former CEO to accept a position as a subordinate if the new boss is 
unambiguously more experienced, or at least more experienced in a crucial way. Even so, the more 
different the successor-CEO's fimctional background is from the founder-CEO's, the more likely the 

founder-CEO's exit from the startup, heightening the need for boards and successors to focus on 
facilitating a smooth post-succession transition. 

If the founder-CEO was fired because of the startup's success, that success can make the startup 
more attractive to new CEO candidates. Henry McCance of Greylock explained how Wily's success 
affected the CEO search: "As Wily hit more milestones, the team realized they could shoot for a more 
capable and experienced CEO." At the same time, if the replaced founder's prior success makes him 
or her unhappy with the change, this can cause problems for the new CEO, leading many successors 
to prefer losing the "irreplaceable founder" rather than retaining a disgruntled founder in an important 
role. 

AFTER THE SUCCESSION 

Even after the startup has found its new CEO, much hard work and many land mines remain for the 
replaced founder-CEO, the incoming professional CEO, and the board of directors. Decisions made 
at this point can still make or break the transition and spoil the startup's prospects. 

"A Tidal Change" 

Even when founder-CEOs have been convinced to support — or at least not actively oppose — their 



own succession, they are ill prepared for the jarring change it introduces. As professional CEO Les 
Trachtman observed, "The founder-CEO says, 'Okay, I'm ready for this,' but they don't have a clue 
what it means. They think it's a title change, not the tidal change it really is." This can be true even 
when the founder-CEO was the one who initiated the succession. Mike Brody at Transcentive, 
despite being fully on board and supportive of Les taking over as CEO, still e?q)erienced a rocky 
transition. "Mike really didn't know what this change meant," e?q)lained Les. To conplicate matters, 
Mike found it hard to let go of day-to-day decision making, and his role on the board signaled to 
enployees that he was still in charge. "He was still the person that people went to when they had 
problems. It was still, 'Mike, may I?' — ^which dis-enpowered me a bit," reflected Les. 

Before joining the startup, a CEO candidate needs to assess whether the startup and the founder are 
ready for such a drastic change. At the startup level, is the startup at the inflection point where a new 
CEO's skills are needed and he or she can have an immediate and positive impact? For instance, is 
the startup heading into a potential "market chasm" that the new CEO's skills will help it cross, as 
was the case for Les Trachtman at Transcentive, or is the startup still facing early technical 
challenges? (The latter was the case when Les failed at Megaserver, the first startup into which he 
was hired as professional CEO.) Will there be fierce opposition to an outsider taking over, or do the 
key employees understand the need for change and see a new CEO as the solution to the startup's 
problems? Does the board fully support the incoming CEO, or are there sensitivities and constraints 
that the new CEO needs to consider when making strategic, cultural, and personnel decisions? At the 
individual level, is the outgoing founder-CEO control-motivated and thus likely to vehemently resist 
the change, or is he or she wealth-motivated and more likely to see the value to be added by a new 
CEO? Is the founder still intensely driven, or is he or she burned out? If the outgoing founder-CEO 
does not fully support the succession, will his or her loyalists also resist it? Or do they, despite their 
loyalty to the founder-CEO, see the need for change better than he or she does? 

In any company, it is an uphill battle for a new outside CEO to learn how the organization operates, 
develop relationships with the key players (many of whom might be unhappy with the change), and 
take hold of the reins. Startups typically add two more complexities to this challenge. First, the new 
CEO is being hired specifically to make dramatic changes in a fast-paced environment, which both 
denies him or her the luxury of taking a long time to get to know the organization and its people and 
also makes resistance more likely. Second, the replaced founder-CEO, in many cases already 
disillusioned with the change, is usually kept on in a prominent position. 

The Replaced Founder's New Position 

In large conpanies, the replaced CEO almost never remains on the executive team and generally 
leaves the board of directors as well. This is invaluable for helping the new CEO take charge and for 
signaling that "there is a new sheriff in town." In most startups, though, the replaced founder-CEO 
remains on the executive team and on the board. In my dataset, even when the board initiated the 
succession, 63% of the replaced founders remained in the startup as executives; when the founder- 
CEO initiated the change, 76% remained in the startup as executives. 

The trigger of change had an even more dramatic inpact on whether the founder remained on the 
board and on which sub-CEO position he or she assumed. At the board level, if the founder-CEO 
initiated the change, he or she remained on the board 96% of the time; if the board initiated the 
change, that number dropped to 60%. Either way, in a majority of startups, the new CEO had to deal 



with having the replaced founder on the board of directors. 

Some boards try to smooth the founder-CEO's transition by letting him or her become or remain 
chairman of the board of directors. Clearly, this can cause problems for the incoming CEO who wants 
to send a clear message about who is in charge and wants to sideline any potential sources of 
resistance to the upcoming changes. At Wily, Richard Williams would replace Lew Cirne as CEO 
only if Lew also stepped down as chairman of the board. Lew recalled Richard's words: "Lewis, I 
think the world of you, but you're not the right person to be chairman. I want David [Strohm, the VC] 
to become chairman instead. I'll only take the CEO job if he is the chairman." From Richard's point 
of view, this would send a clear signal that he, not Lew, was now in charge of Wily. But for Lew, this 
was close to the last straw. "With my president and CEO titles going away, being chairman takes on 
more importance with my knowing that I still have a role that goes beyond technology leadership." 

At the executive level. Figure 10.3 shows that if the founder initiates the change and wants to 
remain in the startup, he or she is almost guaranteed to receive a C-level position, but if the board 
initiates the change, the chance of receiving a C-level position plummets. Founder-CEOs with 
technical backgrounds are most likely to move into the CTO or VP- engineering positions; founder- 
CEOs with business backgrounds are most likely to move into the VP-business development position, 
followed by the VP-marketing position. 

Even though the predominant pattern is for the percentage of founders in executive positions to 
decrease as the startup ages (due to founder attrition and the hiring of nonfounders), there is a distinct 
temporary uptick in the percentage of founders in sub-CEO positions as former CEOs are moved 
downward into those positions. For instance, before any outside financing is raised, only 12% of VPs 
of business development are founders, but that jumps to 38% after the first round as replaced founder- 
CEOs are moved into that position. Before raising financing, 43% of CTOs are founders, but that 
jumps to 57%) after the first round as replaced founder-CEOs are moved into that position; the 
percentage increases again slightly after the second round. 



60% 



□ Trigger of change: Board 



■Trigger of change: Founder-CEO 




Moved to CTO 
orCSO 



26% 



24% 



Moved to 
other C-level 
position 



25% 



49% 



Moved to 
lo>Mer-level 
executive role 



13% 



Left the 
company 
immediately 

37% 



2% 



24% 



Figure 10.3. Post-Succession Positions of Replaced Founder-CEOs 



For Lew Cirne, who was replaced when Wily had a relatively complete organizational structure, 
such a redeployment to a lower-level position could have been even more disruptive had it required 
him to displace one of the loyal executives he himself had hired. "It was a real challenge," Lew said. 
"We had a VP of engineering, a chief scientist, a VP of product management. I had given them 
responsibility and I didn't want to now take that back. Everyone was very cognizant of, 'Where does 
Lew fit?' " Wily had not hired a CTO, however, so Lew was given that title. Yet even this decision 
was not as harmless as it might seem. When Lew saw the new org chart, he realized that his role as 
chief technology officer would be symbolic at best; with no direct reports, he would be "chief of no 
one. This could certainly have been taken as an insult and have fiarther provoked Lew to resist the 
new CEO. 

Keeping the Founder: High Risk, High Return? 

By keeping Lew around, Richard Williams gained the benefit of his insights, institutional knowledge, 
and customer relationships, as well as greater buy-in fi-om the employees Lew had hired. Had Lew 
left before Richard came on board, the startup would have suffered a setback in enployee morale and 
technical direction. As a VC observed, "You can replace a CEO, but you can't replace a founder." 

On the other hand, keeping the founder around can be risky, especially if the founder is control- 
motivated, unhappy with the change, or both. The transitions at Segway, maker of the two-wheel 
electric vehicle, highlight the perils of keeping the founder around when changes are needed but the 
founder does not want to give up control. Early on, founder Dean Kamen decided that his new 
conpany needed a strong and experienced leader — someone who could lead the product- 
development engineers, leverage knowledge of the automotive industry, and bring the product to 
market quickly. He wanted someone who could realize the product's enormous, game-changing 

potential and who "knew what was required to make a global business."^ ^ Through one of his board 
members. Dean found Tim Adams, then a senior VP at Chrysler and "responsible in large part for the 
conpany's fabled turnaround." Dean convinced Tim to join Segway with assurances that Dean 
wouldn't micromanage the company. However, soon after Tim came on board. Dean began to find 
fault: "You talk to [Tim] and he says a few bright things, but he says the same few things every time. 
He's a sinple guy in a lot of ways. . . .He's not nearly as bright as most of the engineers I have, even 

though I pay him two or three times as much as any of them"^-^ Dean placed restrictions on the terms 
that Tim could offer prospective employees and suppliers, dragged his feet on permission to build a 
factory, and prohibited test-marketing in order to preserve secrecy. Dean was the only one who could 
grant clearance to see the prototypes. Eventually, Dean replaced Tim and proceeded to hire and fire 
CEOs nearly annually, severely impacting the growth of Segway over the next decade. 

The challenges faced by a startup's first nonfounding CEO are steep enough that a VC once told 
me, "When I search for a CEO to replace the founder, I usually have to find two CEOs because the 
first one often fails." Similarly, another investor compared the first nonfounding CEO to "a new organ 
being transplanted into a body that will try to reject it." In Figure 10. 1, we can see that by the time of 
their D-rounds of financing, 23% of the startups have had a least two nonfounding CEOs. 

The risk that a professional CEO might fail (or simply decide to leave) fiirther conplicates the 
question of whether to keep the former founder-CEO on board. The founder can serve as an insurance 
policy: In case the body does reject its new organ, the board can hand over the reins to an 
experienced hand rather than suffering a void at the top. 



Smoothing the Transition: The Board's Role 

As we have discussed, it is natural for a founder-CEO, especially a successful one, to resist his or 
her own succession. Such resistance can be extremely disruptive — even dangerous — for the startup. 
For this reason, many professional investors, as part of their due diligence before they initially invest, 
try to gauge whether the founder-CEO will resist an investor-driven succession when the time comes. 
In what one VC referred to as a "rich-versus-ldng test," they assess whether the founder-CEO is 
motivated by control of the startup or by financial gains. If the former, many VCs will shy away fi^om 
making the investment, e?q)ecting the founder-CEO to resist succession should it become (in their 
eyes) necessary, a resistance that could bring down the startup or at least substantially reduce its 
value. (At the very least, the succession is likely to be very tense and disruptive.) If the latter, the VCs 
can be more confident that the founder-CEO's interests will be aligned with their own (financial) 
interests and that he or she will be more likely to agree with them on the need for a new CEO who 
can better advance their mutual interests (although even the most financially motivated founders often 
hesitate when the time comes to step down). 

Investors who have the tough pre-investment conversation with the founder — laying out their 
succession expectations before emotions and tension are heightened — can often lay the groundwork 
for a smoother transition when the time comes. Leaving such a tension-filled discussion for the last 
minute tends to increase founder resistance and organizational disruption. In addition, investors — 
once they have invested — should make sure that the board conducts regular, candid, written reviews 
of the founder-CEO's performance to help provide a flow of developmental feedback, building trust 
and reducing the risk of unpleasant surprises. 

When the time comes to make a change, if the investors have already and correctly diagnosed the 
founder as wealth-motivated, they should highlight for him or her precisely what the new CEO will 
be able to do to build the startup's value that the founder cannot do; for example, the professional 
CEO has a deep sales background that will be immediately useftil and which the founder lacks, the 
professional CEO has already successfully managed a conplex startup whereas the founder has not 
managed anything bigger than a project team, or the professional CEO will be able to upgrade the 
executive team while the founder is held back by longstanding personal relationships with team 
members. Les Trachtman, for exanple, was able to remove longstanding employees and founders and 
cut fi"ee fi"om the family culture at Transcentive, which founder-CEO Mike Brody had been unwilling 
to do. Les was also able to enhance many of Transcentive's business processes — including a 
conplete reorganization of the sales-and-marketing function, which resulted in sales growth of 40% 
per year — an acconplishment that was well beyond Mike's skills. 

As we discussed above, founders who can't see much difference between their own capabilities 
and those of their potential replacements will find it harder to support their own succession, but 
wealth-motivated founders who see distinct, material differences will be more likely to do so. Lew 
Cime's successor at Wily, Richard Williams, had 30 years of high-level experience working at 
conpanies such as IBM and Novell. He had been VP in a variety of positions, including VP of sales, 
and had led two turnarounds and several successful exits. Lew could therefore admit that Williams 
could be a "vitally important leader of the company. We had different things we could bring to the 
table and we respected and worked with that." In fact. Lew felt that helping recruit someone of 
Richard's caliber was one of the most valuable things he had done for Wily. "What made me 
comfortable about getting on board with [the change]," Lew explained, was the knowledge that, as 
one investor said, "[a] person like Dick Williams would not have the know-how or the interest in 



being a CEO of a startup, doing what Lew did." 

As we have seen, many founders don't appreciate the dramatic changes that occur between the 
different stages of a startup's growth. The board may be able to make the founder-CEO more 
supportive of succession by helping him or her understand what those changes will be. Once Lew got 
over his initial shock of being replaced, his mentors were able to help him see how different the 
startup was becoming: "What helped me get through the shock was the realization that the world's 
best speedboat captain isn't able to pilot an oil tanker. It's not who's the best flat-out leader, but 
who's best suited to the tasks at that stage of development." 

Even if the founder is not primarily wealth-motivated, boards can point to other ways in which the 
transition is consistent with the founder's motivations. If the founder loves to plot technical direction 
or lead scientific tasks, but has found the business aspects of the CEO position less enjoyable, boards 
can highlight the advantage of bringing in someone else to take those tasks off the founder's hands so 
he or she can concentrate on more satisfying work. If the founder aspires to build a long-term, multi- 
startup career as a serial entrepreneur, the board can highlight how a smooth transition can support 
rather than hinder that aspiration. Recall that Lew Cirne had aspired "to grow professionally by 
achieving as a businessperson, not as a technologist" even before founding Wily. As CEO of Wily, 
Lew was building his business skills gradually through trial and error. However, the board could 
have emphasized that by accepting a position under a very experienced CEO, Lew would be able to 
"learn from the best" and gain a more solid grounding in the business skills he lacked. Shifting Lew's 
perspective from his immediate losses to his potential longer-term gains could have been a crucial 
step toward getting his support. The argument that Richard could increase the value of Lew's equity 
share, which he could then invest in his next startup, would also help Lew see how a smooth 
transition could boost his longer-term career. 

Rather than forcing an abrupt and permanent fransition, some boards use an interim approach to 
smooth the transition. Lynx Solutions hired Clark Evans as a "bridge CEO" who would be at the 
conpany for only 18 months. As founder James Milmo explained, Clark wanted to help the founders 
"bring stability to a small fast-growing company, then leave to take on the next one." Clark infroduced 
formal processes and sfructures and helped change the sfrategic direction. Clark's status as an interim 
CEO made him less threatening to the founders and helped alleviate their tensions concerning loss of 
control, while giving the startup the new skills it would need for its next stage of development. 
Another interim approach is a two-stage transition, such as Les Trachtman's at Transcentive, wherein 
a potential CEO is hired as a non-CEO executive who might eventually take over as CEO once the 
founder, the potential successor, the board, and the rest of the startup become more comfortable with 
such a transition. 

Gaming the Predecessor's Support: The New CEO's Actions 

The new CEO can also play a critical role in gaining the support of his or her predecessor. When 
taking over from Mike Brody at Transcentive, Les Trachtman gained Mike's support by paying close 
attention to the Three Rs that affect the tensions within a startup team: 

• Relationships. Les spent considerable time getting to know the founder, the conpany, and the 
enployees. His engagement with Transcentive began when he was a consultant, then 
deepened when Mike asked him to become CEO. But before taking on that role, Les insisted 
on taking the role of VP of operations. This helped him evaluate the situation before deciding 



whether to accept the CEO position and gave him a solid foundation for taking charge once he 
did accept it. By the time Les had taken over as CEO, he and Mike had a productive working 
relationship, Mike respected Les's abilities, and they had developed trust in each other's 
motivations. (This was quite a contrast with Les's first startup as a professional CEO, 
Megaserver, when he dove in and began taking action with no prior working relationship with 
the chairwoman or the former founder-CEO.) E?q)erienced facilitators can also help founders 
and successors develop such trust. 

• Roles. After taking over as CEO, Les seized opportunities to educate Mike about his new non- 
CEO role, gently moving him away fi"om the day-to-day functions. "Mike came to me a few 
weeks after I became CEO," Les recalled. "He said to me, 'I'm concerned about how much 
we're spending on milk for enployees.' I told him, 'Mike, I see my job as fi"eeing you up to 
focus on things that really count. This isn't one of them From now on, if it's not a $100,000 
problem, don't spend your time on it' " These incidents helped Mike begin grappling with the 
"tidal" change in small ways, which over time created a very different relationship between 
the founder and the startup with which he had been so deeply involved. The most effective 
approach is a blend of keeping the founder out of tasks he or she should no longer be involved 
with and involving him or her in targeted decisions. Sometimes, the founder can take over and 
be accountable for a specific function for which he or she is particularly suited. At Wily, for 
instance, Richard Williams kept Lew engaged with hiring, which helped Lew feel more 
valued as part of the startup while providing Richard with insights about how each candidate 
would fit into Wily's culture. At other times, the founder might be better suited to a "special 
projects" role that falls outside of the startup's functional boundaries. 

• Rewards. Perhaps one of the most important actions that Les took was to help Mike gain 
significant wealth at the same time that he had to give up control. Les negotiated a partial sale 
of the business to two strategic partners, but used most of the proceeds to cash out the 
founders and early employees who had recently been fired as part of the transition. Les 
explained how this "partial founder buyout" affected both him and his predecessor: "The 
partial buyout acted as a pressure-release valve for Mike. He heaved a big sigh of 
relief . . . and didn't care as much about what happened after that. Abruptly, he said he was 
moving to Virginia and was no longer going to participate in day-to-day. For me, it was a 
different kind of release. I had effectively gotten rid of Mike's brother and the three quasi- 
founders, but now that they had something tangible to take away from it, that made me feel 
great!" In other startups, the board may provide financial incentives for a smooth transition by 
giving the founder a "CEO transition" bonus or by awarding the founder additional equity that 
vests. 

The particulars will be different in every case, but by paying close attention to each of the Three 
Rs, professional CEOs can help their boards smooth the founder-CEO succession, decrease the risk 
of a failed transition, and increase the chances that the replaced founder will remain productively 
engaged with the startup he or she founded. 

CLOSING REMARKS 

Both boards and founders should see founder-CEO succession not as an event but as a process. As an 
event, it is likely to be traumatic, but as a process, it has more opportunity to be productive and 



rewarding as well. From the board's perspective, the process begins even before investors 
participate in their first round of financing, when they have to assess whether the founder-CEO's 
motivation will hinder or facilitate a smooth transition, should the need for one arise. At that point, 
the investors should openly discuss the possibility of succession and the conditions that might trigger 
it* The process continues with the board's ongoing coaching of the founder-CEO as he or she 
encounters new challenges and with new hires who can bolster the founder-CEO where he or she has 
weaknesses. As the process continues, it may come to the point where the board can no longer put 
Band-Aids on the founder-CEO's deficiencies and decides to replace him or her. Bob Davoli, a 
venture capitalist at Sigma Partners, describes this progression: (a) Let the CEO run the company; (b) 
when a problem occurs, the board should identify it and try to work with the CEO to solve it; (c) if 
that doesn't work, "fire him"^"^ Other VCs describe this approach as "coach, then replace." 

The process should also include frank dialogue about the startup's upcoming challenges and 
whether the board has confidence that the founder-CEO will be able to tackle them At Wily, board 
member and investor David Strohm made it clear from the outset that Lew would eventually be 
replaced. "Our initial . . . investment and my involvement in Wily had been predicated on an 
understanding that we would need to bring in an experienced CEO." Even so, confident, passionate 
founder-CEOs often fail to receive that message, especially when they are succeeding. Lew Cirne, for 
his part, was later blindsided by David's explicit request that he step down, describing his reaction 
as "shock." Clearly, he either had not heard or had not grasped David's earlier assertion that he 
would have to give up leading Wily relatively soon. Henry McCance, the chairman of Greylock, 
observed, "An entrepreneur who is anxious to receive capital, may say, 'Okay, sure, I will do 
whatever it takes.' But he may be thinking, 'I'll show them that I can do it all!' " Even when directors 
think they have sent the message about succession, they must realize that, most likely, the message was 
not understood clearly, not taken as seriously as it was meant, or tuned out altogether. 

Founders, by acknowledging succession as a process that starts early in the startup's life, can at 
least gain more control over how and when they lose control of their startups. As we will see in 
Chapter 11, even before founding their startups, founders should reflect deeply on their core 
motivations and then make founding decisions that are consistent with those motivations. If those early 
decisions include taking capital from outside investors, the founder must understand that he or she 
may already have taken a key step toward his or her own succession. 

Founders should also realize that they can gain some control over when they lose control of the 
startup by "getting ahead of the board" and initiating the succession themselves.* Founder- initiated 
succession and board- initiated succession can have very different outcomes. As described above, 
when the founder triggers the change, he or she is less likely to leave the startup immediately, more 
likely to play a central role in the search for and choice of a successor, more likely to remain in a 
senior executive role, and more likely to remain on the board of directors. Triggering one's own 
demise as CEO is wrenching but, for some founders, is worth the price if it can ensure their remaining 
a prince of the realm after relinquishing the throne. 



* In all, 62% were triggered by the board. Other triggers of succession — such as investors who were not on the board, a recent 
executive hire, or even a non-CEO cofounder — accounted for the remaining 11%. Note that this does not include unsuccessful attempts 
to fire the founder-CEO. Field research suggests that such attempts are not uncommon, but has not yet established systematically how 
frequently they occur or how founder-CEOs survive them. 



* Sometimes that self-awareness is enhanced by participation in regular meetings of a CEO forum or by working with an outside 
"CEO coach" who becomes a trusted, unbiased advisor. Participation in a forum can also mitigate the loneliness of being a CEO, which 
Dick Costolo described as being "alone at the center of an hourglass." 

* Some startups later face a second dramatic shift in their customers and their market focus when they try to "cross the chasm" 
(Moore, 2002). Companies developing products that are innovative or based on new architectures (or both) have to transition from an 
early market dominated by a few earfy adopters to a larger mainstream market. The chasm such companies must cross includes a huge 
difference in the skills and operations required to serve these different markets, a huge difference in customer preferences, and other key 
issues that involve multiple functional areas. Although the challenges of crossing the chasm are classically seen as a marketing problem, 
they also require new capabilities at the top of the startup and thus often require a new CEO. Far from being just a marketing issue, the 
challenges of crossing the chasm should be seen more broadly as a leadership and operational-management issue. 

* In a different realm, how often has the right person to lead a revolution proved to be the wrong person to subsequently lead a 
govenunent? For instance, long ago, in ancient biblical history. King David was a warrior king who led the Israelites during the conquest 
of their country and their capital city, Jerusalem, but was deemed unfit to establish the Great Temple and other administrative features of 
a stable kingdom. Instead, his successor, Solomon, was far more suited to carrying out those tasks, and did so successfully during his 40- 
year reign. 

^ See Chapter 9 for more details. 

* See Chapter 4 for more details about homophily and its powerful impacts on team composition. 

* If investors do not raise the issue, founders should not assume that this means they will never be replaced. Even though it is natural 
for a founder either to be optimistic about the probability of remaining CEO throughout the life of the startup or to want to avoid a tense 
discussion about succession possibilities, founders who are concerned about being replaced (and who might think twice about taking 
capital from investors who might want to replace them) should fight those inclinations and raise the issue before accepting the capital 

* In contrast, VC Jeff Bussgang observes, "If your board is 'ahead' of you on CEO succession, it's a problem." 



PART IV 
CONCLUSION 



CHAPTER ELEVEN 



WEALTH-VERSUS-CONTROL DILEMMAS 



The path from founding to success is a long and winding one, with dilemma after dilemma 
forcing founders to make decision after decision, all with important — and sometimes surprising — 
short-term and long-term consequences. Throughout this book, we have examined a wide variety of 
players — ranging from core founders to cofounders, hires, investors, and successors — and a wide 
variety of dilemmas. Those who have made the decision to found a high-potential startup must decide 
whether to go it alone or to assemble a founding team. Those who choose to assemble a team need to 
make decisions about the relationships, roles, and rewards of their cofounders. Both solo founders 
and founding teams are likely to have to make decisions about the relationships, roles, and rewards of 
their hires. For many founders, there will be further decisions about the various kinds of investors. 

We have also seen a wide variety of outcomes that result from those early decisions. Sometimes, 
they were what the founders had hoped from the start to achieve. However, we have also seen many 
ways in which things don't work out as the founders want, from Kaleil and Tom having to pay off 
their former govWorks.com co-founder, Chieh, to Lynx cofounders James Milmo and Javier Pascal 
facing the prospect of having to sell their startup for only enough money to pay back their investors — 
not themselves — to Lew Cirne being forced out of his CEO position at what appeared to be the height 
of his success. 

We have looked at each of these outcomes in light of the decision points that led up to them We are 
now in a position to understand them collectively in yet another light — a "wealth versus control" 
dilemma that coexists with all the other dilemmas (once one has decided to found at all). But first, 
let's step back and once more look at the basics. 

What is entrepreneurship? A widely used definition is "a process by which individuals pursue 
opportunities without regard to the resources they currently control."^ That sounds sfraightforward, 
even romantic, but it has a dark side: When a founder of a high-potential startup chooses to pursue an 
opportunity regardless of whether or not he or she has the necessary resources, a critical piece is 
usually missing — often several critical pieces. One study estimated that founders are 60 times more 
likely to be resource constrained than to have all of the resources they need.^ Yet, pursuing an 
opportunity requires having a full complement of the relevant human, social, and financial 

resources.^ The more resources a startup can control and the more quickly it can gain control of them, 
the better its competitive position.^ Lacking resources is a big reason why new ventures have such a 

high failure rate.^ Lack of resources lies behind all the dilemmas we have described so far. If a 
founder started out with all the human, social, and financial resources he or she needed, there would 
be no need for cofounders, hires, or investors, nor any need for him or her to be replaced as CEO. 

Attracting outside resources, especially those in short supply,^ requires founders to give up 
valuable assets.^ Outside resource providers often want two major things in return for their 
contributions to building the value of a startup: economic ownership and decision-making control. 



Regarding ownership, resource providers are often motivated to contribute their resources by the 
chance to gain a share of the economic winnings should the startup succeed. coft)unders and hires who 
know their own value want to have a hand in important decisions in return for the skills and contacts 
they contribute; investors also demand a degree of control over the organization's decision making in 
return for their guidance and capital.^ Control of a startup can be contested in two ways: whether the 
entrepreneur remains CEO or is replaced,^ and whether outsiders or insiders control the board of 
directors. Founders who reftise to give up ownership and control in either or both of these ways 
will be less likely to attract the resources they need and thus not be able to ftilly pursue the 
opportunities they envision. It appears, then, that each of our founding dilemmas is also a dilemma of 
what resources to acquire at what cost in ownership and control. This is the dilemma behind all the 
other dilemmas. As we will see, it can be the harshest of the dilemmas in this book. 

A CENTRAL DILEMMA: WEALTH VERSUS CONTROL 

It is harsh because it pits the two most common entrepreneurial motivations — wealth and control — 
against each other.* Most founders embark on their entrepreneurial journeys with passion and 
confidence, expecting to build a valuable startup and then run it throughout its life. They want to be 
the next Bill Gates of Microsoft or the next Anita Roddick of The Body Shop, prominent founders 
who managed to achieve both value creation and control. However, few founders are able to 
maximize both goals because, at every stage of startup evolution, the actions that maximize one 
inherently hinder the other, and most founders are forced to choose between one or the other of these 
goals. In short, there is an inherent tension between achieving one and the other. (While other 
entrepreneurial motivations can be important, they lack this inherent tension.) 

To show this pattern more clearly. Figure 11.1 summarizes some of the key dilemmas already 
covered in Parts II and III of this book, but now reveals them to be a sequence of necessary trade-offs 
between maintaining control and creating a valuable startup or, as we noted above, a recurring 
dilemma of what resources to acquire at what cost in ownership and control. As discussed below, the 
decisions listed in each column are the decisions that each type of founder should make, given a goal 
of either maximizing personal control or maximizing the eventual payoff. In real life, the choices are 
seldom as consistently aligned as they should be. 



Putentul 



PartiiifKltits 
til the HtJrllip 


Dt\ iiiim Arcj 


Decisions OrittitfJ tmi'jrj 
SljintJimng Coutml 


riecisions Oriented tnu-.irJ 
Miixiniixing Weitllh 


Cofoundcrs 


Solo vs. team 
Relationships 

Roles 

Rewards 


Remain solo founder (or attract 

weak cofoundcrs) 
First look to immediate circle for 

"comfortable" cofounders 
Keep strong; control of decision 

making; build hierarchy 

Maintain most or all eqiiits' 
o>«'nership 


Build founding team; attract best 

cofounders 
Tap strong and weik ties to find the 

Isest (and complementary) cofounders 
Give decision-making control to 

cofounders with expertise in specific 

areas 

Share equity to attract and/or motivate 
cofounders 


Hircv 


Relationships 

Roles 
Rewards 


Hire ss'ithin ckwc personal network 
(friends, family, and others) 
as required 

Keep control of key decisions 

Hire less expensive junior emploj-ees 


.•\g^;res>ively tap broader network 
(unfamiliar candidates) to find the best 
hires 

Delegate decision making to appropriate 
expert 

Hire e.\pericnced employees and incent 
them with cash and equity 


Investors 


Self -fund vs. take 
outside capital 
Sources of capital 


Self-fund; "bootstrap" 

Friends and family or money-only 
angels; tap alternative sources (e.g., 
customer prepayment* or debt) if 
possible 


Take outside capital 

Target experienced angels or venture 
capitalists 




Terms 

Board of directors 


Resist investor-friendly terms (e.g., 
refuse any supermajority rights) 

.Avoid building ofiicial board; when 
built, control composition and 
makeup 


Be open to terms necessary to attraa best 
investors (e.g., supermajority rights) 

Be open to losing control of board if 
necessary to get best investors and 
directors 


Successors 


Trigger of succession 

OpeiuKss CO succession 

Desired role after 
succession 


Avoid succession issue until forced 

Resist giving up the CEO po«if ion 

Prefer to leave than to remain 
"prince" 


Be open to initiatmg succession when 

next sMge of startup is outside one's 

own expertise 
Be open to giving up CEO position to 

better CEO 
Want to remain executive in position 

that matches skilb and preferences 


Other (actors 


Preferred rate of startup 

g/o\s'th 
Capital intcntity 
Core founder's "capitals* 


Gradual to nuxlcratc 

Low capital intensity 
Well equipped to launch and build 
startup without mixh help 


Fast to explosive 

High capital intensity 
Important gaps that should be filled 
by involving others 


Mmt likclir 
outcome 




Maintain control; build less valiK 


HuiUI linarKial value; imperil OKiirol 



Figure 11.1. Wealth-versus-Control Dilemmas 

The decisions in the first column consistently prioritize maintaining control over building financial 
value. Even when a potential co-founder, hire, investor, or successor would add important human 
capital, social capital, or financial capital to a startup and enable it to build more value, a founder 
who consistently makes control decisions will forgo that extra value in order to maintain control of 
the startup. Founders who consistently make control decisions (a) remain solo founders or choose 
only cofounders who will allow the founder to retain control, (b) hire inexperienced people and keep 
control of decision making, (c) self- fund or raise capital only fi^om investors who won't interfere with 
the founder's control of the startup, and (d) choose to remain CEO throughout all stages of startup 
evolution. Such founders take full responsibility for development and implementation of their visions 
and aim to rely largely on their own human, social, and financial capital. The data (see below) 
confirm that such founders are more likely to maintain control for a longer period of time but end up 
with less- valuable equity stakes. 

Conversely, decisions in the second column consistently prioritize building value over maintaining 
control. This second type of founder (a) strives to attract cofounders whose expertise fills important 



holes, (b) hires experienced people who take control of their domains, (c) raises money from 
investors who add enough value to the startup to justify the control they demand over decisions, and 
(d) watches for the point when someone else will be able to do a better job leading the startup during 
its next stages of development. If a potential cofounder, hire, investor, or successor would add an 
important piece to the startup (skills and human capital, contacts and social capital, or financial 
capital), the founder who is making wealth choices is willing to do what it takes to attract that person, 
even if doing so imperils the founder's own control. 

Many founders, especially first-timers who haven't yet experienced the outcomes of their decisions 
or do not have an experienced mentor to guide them, often do not realize that their early decisions are 
leading them in one or the other of these directions. Unfortunately, Figure 11.1 is only a normative 
description of the systematic decisions that different founders should make — ^but often do not. As we 
have seen throughout this book, systematic consideration of early decisions can yield much better 
outcomes than an accumulation of reactive decisions; founders need to "decide rather than default." 

Divergent Outcomes: Rich versus King 

At each fork in the road, the decision that maximizes value tends to threaten the founder's control, and 
vice versa. There is an inherent confiict between maintaining control and building value in high- 
potential startups because the latter requires value-added players who demand some control. 

Founders who consistently make control decisions are more likely to reach what I call the "King" 
outcome, in which the founder retains the throne but does not rule as big and rich a kingdom as might 
otherwise have been possible. Founders who consistently make wealth decisions are more likely to 
reach what I call the "Rich" outcome, in which the founder generally loses the throne but sees his or 
her venture pursue its business opportunity to the fiillest. Figure 1 1.2 summarizes these two outcomes, 
the outcome of achieving both control and value creation ("Rich and King," also known as "the 
entrepreneurial ideal") and the outcome of achieving neither control nor value creation ("Failure"). 

My quantitative analyses suggest that few founders — especially first-time founders — can maintain 
control and create maximal value.* As we saw in Chapter 10 on founder-CEO succession, few 
founders — even the exttemely successfiil ones — are still CEO when their startups become very big or 
go public. Bill Gates and Anita Roddick are so well known precisely because they are the 
exceptions. Yet, many founders have so much confidence in their startup's prospects and in their own 
abilities and feel such intense passion for their idea that they systematically underestimate their need 
for ftjrther resources or overestimate their ability to remain in conttol of the startup even as they take 
on a wide range of outside resources. These natural entrepreneurial inclinations can lead founders 
away from becoming Rich or King as they make decisions that leave them with either too few 
resources or too little control of the startup.* 

Financial Gain 

Far Below Achieving 
Potential Value Potential Value 

Failure Rich 
King Rich &c King 



Retention of Minor player 

decision-making Major player 
control 



Figure 11.2. Rich versus King Outcomes 

Adapted and reprinted by permission of Harvard Business Review. From 'The Founder's Dilemma," by Noam Wasserman, February 
2008. Copyright © 2008 by the Harvard Business Publishing Corporation; all rights reserved. 



My analyses also suggest that founders who keep control personally give up a significant amount 
financially. Such founders tend to build a less valuable startup while keeping a larger share of equity 
in it, but it turns out that the value-seeking founder's "smaller slice of a larger pie" is generally 
greater than the control- seeking founder's "larger slice of a smaller pie." Controlling for a wide 

variety of differences across the 460 startups I analyzed/^ founders who had kept control of both the 
CEO position and the board of directors held equity stakes that were only 52% as valuable as those 
held by founders who had given up both the CEO position and control of the board, as shown in 
Figure 11.3. (The founders' equity stakes had an intermediate value if they had given up control of 
either the CEO position or the board but not both.) This result held in both younger startups and older 
ones, suggesting that founders face a trade-off between wealth and control, and that this trade-off 
persists throughout the stages of startup evolution. 




Figure 11.3. Value of Founder's Equity Stake, Depending on Degree of Control Maintained Reprinted from "Rich vs. King: The 
Entrepreneur's Dilemma," by Noam Wasserman, Best Paper Proceedings of the Academy of Management, 2006. 



To assess whether founders tend to make consistent decisions, I compared solo founders (whom 
we would expect to be more inclined toward control) to the core founders who attracted cofounders.* 
This dataset included 1,658 technology startups that completed surveys between 2005 and 2009, and 
it controlled for a wide variety of founder and startup characteristics, such as capital intensity, 
industry segment, and startup maturity. The statistically significant results indicated some consistency 
that extended across our founding dilemmas, as summarized below: 

• Hires — Solo founders hired younger employees than multi-founder startups did, suggesting 
that solo founders favored more inexpensive hires. 

• Investors — Solo founders (a) were almost twice as likely to use debt, which does not dilute 
the founder's ownership, as a source of financing; (b) raised their first institutional round of 
financing later in the lives of their startups (possibly because investors were more hesitant to 
invest or possibly because the founders wanted to wait longer); and (c) raised less capital in 



their first institutional round of financing (which again may be due either to investor 
preference or to founder preference). 

• Chairman role — Solo founders were more likely to retain both chairman and CEO titles (i.e., 
remain a "double King") after their first round of financing. 

Solo-founded startups received lower pre-money valuations than did multifounder startups, 
confirming the price that control-motivated founders pay for making control-motivated decisions. 
However, the results above also suggest that, while solo founders raised less capital, they may have 
done so on more fi-iendly terms, a trade-off that control-motivated founders are usually willing to 
make. 

Using the Wealth-versus-Control Lens to Revisit Evan Williams's Decisions 

Throughout this book, we examined the very different choices that Evan Williams made as he was 
founding and building Blogger and Odeo. With the tension between control decisions and wealth 
decisions in mind, let's take one more look at Evan's choices. At Blogger, he consistently made 
control decisions rather than wealth ones. He founded with a former girlfriend, Meg Hourihan, who 
did not have experience in startups or in the industry. He hired young and inexperienced friends to 
build the product at inexpensive salaries, later replacing them with an even more extreme choice — 
volunteers. He was adamant about keeping a controlling interest in the company, refusing to split the 
equity equally with Meg and avoiding selling equity to professional investors. When he disagreed 
with Meg, he overruled her or simply ignored her concerns in favor of pursing his own vision for the 
company. He avoided talking to VC investors, even in the midst of the dot-com funding frenzy, 
preferring to accept funding from fi-iends, family, and angel investors, and even taking donations 
during a public "Server Drive." His board was advisory in nature, offering little guidance. When Meg 
challenged Evan for the CEO position, he refused to give it up, leading her to leave the company. He 
also refused an acquisition offer that threatened his control of the startup, even though that refusal 
meant losing his entire staff In fact, after making this last decision, Evan felt relief: "The next day, 
mixed with the sadness and the loss was an incredible amount of liberation." 

With Odeo, Evan consistently made wealth decisions that threatened his control of the startup. He 
founded with Noah Glass, a business acquaintance who had experience in the podcasting segment, 
and gave Noah both the CEO position and the lion's share of equity. When Evan realized the great 
potential of Odeo's podcasting technology, he steered the startup into a series of decisions designed 
to help Odeo grow quickly. He negotiated with a venerable VC firm to raise $5 million, which he 
used to hire high-level experienced executives, technologists, and sales managers who would drive 
decisions within their functions. After that, Evan's board, which was dominated by representatives 
from his investors, became highly involved in the decision making at Odeo. 

Within each startup, Evan's choices were quite consistent. At Blogger, he made the control 
decisions shown in the left column of Figure 11.1 and ended up with full control, but it was full 
control of a startup that sold for much less than its potential warranted, considering that Blogger was, 
at the time, the tool of choice for a rapidly expanding user base. With Odeo, Evan made the wealth 
decisions shown in the right column of Figure 11.1, choices that increased the chances of growing a 
valuable startup at the cost of much of his own equity and decision- making control. 

The further evolution of Odeo is also quite instructive, for it shows us how Evan tried to balance 
his control inclinations with the desire to pursue big opportunities. As Apple's iTunes gained traction 



in the marketplace, Evan and his team realized that Odeo's prospects were much diminished and 
turned to brainstorming other ideas. Jack Dorsey, an engineer at Odeo, proposed a status-updating 
tool that would allow users to send short text messages instantly to a group of "followers." The team 
decided to develop a prototype, and two weeks later a product called Twitter was born. While Evan 
and his fellow engineers saw the wisdom in switching to developing Twitter, it was much harder to 
convince the VCs on the board. By the time he had gotten to this point, Evan had realized that he 
needed full control of Odeo if he were going to be able to explore and develop Twitter's full 
potential, but that his earlier decisions had caused him to sacrifice that control. To reclaim the 
creative freedom he craved, Evan made the nearly unprecedented decision to put up $3 million of his 
own money to buy out his investors and retake control of his company. A year later, Evan sold 
Odeo's podcasting assets to Sonic Mountain for over $1 million, then spun out Twitter with the idea 
founder. Jack Dorsey, as CEO. 

In 2008, as Twitter's growth began accelerating by adding millions of users per month, Evan again 
felt the need to retake control. With the board's approval, he moved Jack fi-om CEO to chairman and 
took over as CEO. Evan was later quoted as saying, "It's hard and confusing ... I think there's few 
cases in history where the CEO steps down and is also the founder and reports to someone and that 
works. "^^ For his part. Jack said that being removed from his position at Twitter "was like being 
punched in the stomach." ^-^ Along the way, Evan was careful to raise capital only at times and on 
terms that were compatible with his desire to retain control. But Evan struggled; in his words, he and 
his team were "just hanging on by our fingernails to a rocket ship."^"^ He couldn't hire or put 
processes in place quickly enough to keep the system running. Twitter also lacked a monetization 
strategy. To fill these holes, in September 2009 he hired Dick Costolo, the former founder-CEO of 
FeedBurner whom we've also gotten to know in these pages, to be his chief operating officer. 
Following the playbook described in Chapter 10, after "trying before buying" for a year, Evan named 
Dick his successor as CEO in September 2010 and moved himself into a product- strategy role before 
transitioning out of Twitter a few months later. Dick Costolo's fifth startup would thus be one in 
which he took over from the founder-CEO, applying all of the lessons he had learned from his first 
four startups to lead someone else's startup through its next stage of development. 

Othe r Alte mative s : Se cond-tie r Re source s 

There are two ways in which Evan's decisions were more extreme than those of many founders: Each 
choice was at an extreme of control versus wealth, and, within each startup, his choices were 
extremely consistent. 

Figure 11.1 presents a series of binary choices: avoiding bringing on a specific high-quality 
resource — ^which may add the most value but may also pose the biggest threat to the founder's control 
— versus bringing on that resource. In real life, however, there are in-between or "second-tier 
resources" that may add more value (but not the maximum) but with less threat (though not without 
any) to the founder's control. Rather than solo founding or attracting the best cofounders, founders can 
choose to attract "adequate" cofounders. Rather than hiring either inexperienced or very experienced 
executives, founders can hire midrange executives. Rather than self- funding or taking money from a 
top VC firm, founders can take money from second- tier investors. Rather than taking no VC money or 
the most they can get, founders can raise smaller rounds of financing and face less risk of being 
replaced as CEO.^^ 



For instance, Frank Addante used second-tier resources for his startup StrongMail when he 
recruited a "B team" of inexperienced people through Craigslist. By using a B team rather than 
expensive top-college grads or experienced executives, Frank was able to lower his burn rate and 
bootstrap the conpany while he worked to gain proof of concept. The downside was that Frank spent 
valuable time having to direct his team and was frustrated by their lack of motivation and self- 
direction. However, the B team fiilfilled its tunction of buying Frank the time that he needed to build 
the company, develop the product, and attract VC attention. He eventually raised outside capital and 
replaced his B team with top-tier hires. 

Other Alternatives: Hybrid Paths 

We have seen that Evan's decisions were remarkably consistent within Blogger and during the early 
stages of Odeo. Each individual decision a founder makes can contribute incrementally to building the 
startup's value or to maintaining control of decision making. We would expect consistent control- 
oriented decisions to markedly increase the founder's chances of maintaining control (though of a 
smaller startup) and consistent wealth-oriented decisions to markedly increase the founder's chances 
of growing more value (while losing control of decision making). Some founders, however, pursue a 
hybrid path — a mix of control and wealth decisions — which they hope will increase the chances of 
achieving both Rich and King.* For instance, they could maintain frill confrol by remaining solo 
founders and by hiring only employees who will not challenge that control. If they are using this 
period of total confrol to develop an initial product and achieve key early milestones, they might have 
more confrol over the terms they get from outside investors. They can then shift sfrategies toward 
making wealth-oriented decisions and fry to "hit the accelerator" by atfracting the best investors they 
can find and leveraging their investors' capital and guidance to scale the startup and the team Hinting 
at such a sfrategy, the chairman of the Buffalo Angel Network was recently quoted as saying, "The 
hardest decision a founder, inventor, or enfrepreneur needs to make is 'when do I give up some 

confrol to grow the company.' "^^ 

At the beginning of her startup, Sittercity, Genevieve Thiers made confrol-oriented choices. She 
chose to be a solo founder; used friends and family connections as advisors, employees, and angel 
investors; and kept all the decision-making confrol and all the equity to herself When Sittercity had 
been in business for eight years, however, she felt the time was right to seek venture capital to 
continue Sittercity's growth and maintain its first-mover advantage. Genevieve explained why she 
waited so long: "A lot of founders take fiinding too early and lose ownership in the company. The 
company starts to be run by the investors. . . . [For me,] being able to retain confrol was very 
important. We finally did decide to take venture capital money in 2008 because ... I always felt that 
— at a certain time — having fiinding would be appropriate for the company. We were probably the 
latest-stage Series A our investors had ever seen. We were so successfiil, so big, and had this 
established team. Our investors could put their money in and just sit back." For startups such as 
Sittercity, with the luxuries of time and low capital-intensity, such a sfrategy can work. For other 
startups, however, such a sfrategy would increase the likelihood of slow or lower-quality product 
development (because fewer resources are being used to develop it) and of conpetitive 
disadvantage, harming the value that can be built and thus the chance of achieving the enfrepreneurial 
ideal. 

In deciding on a wealth, confrol, or hybrid sfrategy, founders should carefiilly assess which of the 



resources most critical to the startup's success are lacking, then carefully allocate equity, financial 
resources, and attention to attracting those resources. Such founders can try to make up for 
weaknesses in one area by bolstering their resources in another area. For instance, a solo founder 
who consciously avoids adding a cofounder (thus retaining all of the equity) may instead focus on 
using equity slices to attract the best post-founding hires. Solo founder Lew Cirne, for example, 
carved out some of his own equity to sweeten the pot for his first employees after his angel investors 
reftised to allow their shares to be diluted. 

Such hybrid strategies may indeed reduce the probability that founders will end up as exclusively 
Rich or exclusively King, but they may also increase the probability of ending up Failures. Just as 
entrepreneurial firms that pursue multiple strategies can get "stuck in the middle," so too should 
entrepreneurs who make inconsistent choices be more likely to end up with muddled strategies'^ and 
face greater risk of failure. For a brief and simple quantitative exploration of the pros and cons of 
hybrid strategies, see Appendix 11-1 at the end of this chapter. The model there suggests that founders 
who are unsure about their core motivations, and thus which outcome and accompanying set of 
choices are best for them, can hedge their bets by using hybrid strategies — ^but at high risk. This will 
markedly reduce the chances of achieving one outcome, slightly increase the chances of achieving the 
other, and significantly increase the chances of failing. 

Another potential risk of hybrid or hedging strategies is that founders who shift strategy in the 
middle of building a startup may cause major tensions within the startup or harm their ability to attract 
the best resources. For instance, if they initially attract the best cofounders and hires (wealth), then 
refuse to raise money from outside investors (control), they may bring on major dissention and high 
turnover among critical employees and lack the resources to continue paying the ones who remain. 
When Dean Kamen — whose medical-product research-and-development company, DEKA, invented 
the Segway personal transporter — first realized Segway's potential, he began talks with scores of 
investors and hired an experienced executive team to manage the project, offering stock options to 
attract the best managers and engineers. Dean's subsequent insistence on control rebuffed potential 
flinders and eventually caused Segway's experienced CEO, Tim Adams, who had been promised 

control of key decisions, to leave the company.'^ 
"BETTER" AND "WORSE" OUTCOMES 

If different decisions lead to such different outcomes, is one outcome better or worse than the other? 
Evan Williams made control-oriented choices at Blogger and then wealth-oriented choices at Odeo — 
was he right in one case and wrong in the other? 

The Core Founder's View: Self-Knowledge and Decision Making 

As we saw in Chapter 2, there is a wide range of entrepreneurial motivations. Many of these 
motivations — altruism, variety, and intellectual challenge, for example — are unlikely to provide 
founders with clear guidance as to which decisions to make. But maintaining power and control over 
the key decisions within the startup and gaining financially from building a valuable startup are the 
two most common motivations, and they do provide clear guidance. 

For the founder, is one set of choices better than the other? In certain circumstances, it may be. 
Given a capital-intensive idea in a ticking-clock environment, for example, control-oriented choices 



will indeed be more problematic than wealth-oriented choices. But otherwise, wealth- oriented 
choices are neither better nor worse than control-oriented choices — as long as they are consistent 
with the founder's core motivation. What's "worse" is a set of choices that are inconsistent with each 
other, which means that some of them must be inconsistent with the founder's core motivation. Worst 
of all would be a consistent set of choices all of which were in opposition to the founder's core 
motivation. 

A wealth-motivated founder who consistently makes control decisions may get to some kind of 
outcome, but it probably won't be the one he or she had hoped for; likewise for a control-motivated 
founder who consistently makes wealth decisions. Early in the founding of Wily Technology, Lew 
Cirne stated a goal of growing as a businessperson by being CEO. To this end, he consistently made 
control choices: founding, working alone for a year to develop the initial product, raising money from 
friends and family, and hiring junior people who might be able to grow. But then, to spark Wily's 
growth, he took multiple rounds of financing from top VCs, lost control of his board, lost his 
Kingship, and was sidelined within his own startup. Lew ended up having a nice payday when Wily 
exited, and a founder whose core motivation is to get rich would have been delighted with Lew's 
outcome. Lew himself, though, regretted having given up control and vowed to keep control in his 
next startup. "Next time," he exclaimed publicly, "I'm running it to a billion dollars, and I don't care 
what any VC says about that!" 

Only by understanding the founder's motivations can we — or the founder himself or herself— judge 
whether the outcome was better or worse. (As discussed below, we also have to consider business 
characteristics or environmental circumstances that might force a founder to go against his or her own 
core motivations.) 

What about Evan Williams? Did he truly have two different motivations as he pursued his two 
different startups? In Blogger, he risked everything to maintain control. In Odeo, he seems to have 
become a different person, willingly trading control for maximization of the startup's value. In fact, 
though, after Evan raised $5 million from VCs, he chafed under the scrutiny of his board of directors. 
"Anything I do," he complained, "I have to explain and justify it to the board. ... It kills the creative 
part that can lead to something good. ... I don't want to be required to explain myself Having a 
board is killing our ability to try new things." He concluded, "I could believe in the company . . . only 
if I had complete control over it." It would seem then that he had talked himself into being wealth- 
oriented when, inside, he was still control-oriented. 

Choosing between the options outlined in Figure 11.1 can be particularly hard for (a) founders who 
lack clear priorities and (b) founders who do not yet know themselves well enough and have not 
experienced the challenges faced by startup CEOs. (Founders who both have unclear priorities and 
lack self-knowledge and experience will be in for a really rough ride.) In the first case, the founders 
may be equally motivated by wealth and control. On the other hand, they may not be that strongly 
motivated by either and thus find it particularly difficult to make coherent decisions. In the second 
case, first- time founder-CEOs are often handicapped by at least four kinds of ignorance. First, having 
never been in a fast-growing startup, they do not understand how the stages of growth will differ. 
Second, they do not realize how those very different stages will cause dramatic changes in the 
challenges faced by the CEO and the executive team Third, they do not yet know whether or not they 
have the skills and capabilities to address those challenges. Fourth, they have not had to refiect on 
whether they will be willing to trade off control versus value creation and are therefore less prepared 
to make decisions that are consistent with their core motivations. 

As a first-time founder. Lew Cirne had little understanding of these four issues. He didn't realize 



that, by building Wily quickly, he would more quickly come to the point at which his experience and 
expertise could no longer fulfill Wily's needs. Lew only gradually came to understand that his career 
as a technical manager had not prepared him for the challenges of leading a maturing company. 
Having never dealt with venture capitalists, he didn't understand that he was risking his control of the 
company by accepting fiinding from them Perhaps most importantly. Lew did not anticipate the 
intensely negative reaction he would have to losing control of Wily. Even though his investors had 
intimated that he would eventually need to be replaced. Lew clearly hadn't foreseen or planned for 
such an event. It was only after much discussion between Lew and his investors, and after more than a 
year of looking for a successor, that Lew realized, "It's not who's the best flat-out leader, but who's 
best suited to the tasks at that stage of development." 

Some founders are self-aware even at the beginning of their first startup. The founder of Steria, an 
IT systems and services company, made decisions that resembled Evan Williams's control-oriented 
decisions at Blogger, but was much clearer from the outset that his primary motivation was "to remain 
independent and master of his own destiny," and he consistently aligned his decisions with that 
motivation. Instead of taking capital from outside investors, he relied on founder capital and bank 
loans; he refiised to grant stock to attract better hires to Steria; and he insisted on remaining CEO. As 
a result, Steria experienced slower growth than it would have had the founder taken outside capital, 
attracted better employees, and been more open to giving some control to experienced executives. 
Each of these latter choices held the promise of greater financial gain, but at the expense of the 
founder's maintaining control, and Steria' s founder avoided them with open eyes.* 

The Views of Other Participants: The Need to Assess the Founder 

Core founders can find it enough of a challenge to make key decisions that are consistent with their 
own motivations. However, as soon as they involve other participants, such as cofounders, hires, and 
investors, the potential for misalignment grows dramatically. Thus, before joining the startup, those 
potential participants also need to assess whether the founder's motivations — and the corresponding 
decisions he or she will make in response to the inevitable sequence of founding dilemmas — may 
conflict with their own motivations. 

Potential cofounders, in particular, need to share similar — or precisely complementary — 
motivations. At each fork in the road, cofounders will have to make a single coherent decision about 
how to proceed. For instance, they will have to decide together if they should raise outside money or 
continue to self-fund. If the cofounders are all wealth-motivated, they are more likely to agree on 
attracting the resources that can grow the value of the startup. A mixture of a control-motivated 
founder-CEO and wealth-motivated cofounders can be harmonious if the control-motivated founder- 
CEO is indeed the right person to be the CEO and will be able to build the startup's value most 
effectively (thus fUlfllling his or her own desire for control and the other cofounders' desire for 
financial gains), but contentious if the wealth-motivated cofounders lack confidence in the founder- 
CEO's ability to build the value of the startup. However, other mixes of cofounder motivation maybe 
far more dangerous. In particular, if the cofounders are both control-motivated, they may end up like 
Evan and Meg at Blogger, with intense startup-endangering fighting over who is CEO and who 
controls the most important decisions. 

Cofounders may have very different ideas about what the startup should do regarding almost every 
dilemma examined here. Such divergences can introduce intense tensions into the team, threatening its 



stability and the startup's growth and survival. Potential cofounders who discuss these dilemmas 
before deciding whether or not to cofound together, walking careMly through each row of Figure 
11.1 and discussing what decisions they would each want to make at each fork in the road, will be 
able to anticipate — and perhaps even avoid — such tensions. They may also realize that they simply 
shouldn't be cofounders to begin with. 

For potential hires, diagnosing the founder-CEO's motivations can also be critical.* Wealth- 
motivated hires should try to work with wealth-motivated founders. If the founder-CEO is control- 
motivated, we would expect the hires to be attracted for other reasons: the founder's compelling 
vision, the chance to "change the world," the excitement of working in a startup, or other nonfmancial 
considerations. Such was the case at Segway, where talented engineers competed to work for Dean 
Kamen, even for below-market compensation. They were attracted to the culture of creativity, the flat 
organizational structure, and the pursuit of the next big idea with a visionary like Dean. "I owe Dean 
for giving me the greatest experience of my life," said one DEKA engineer. 

However, when the founder-CEO starts making control-oriented decisions, such as refusing to 
raise needed capital or to delegate decisions to experienced executives, the wealth-motivated hire is 
likely to chafe under the founder's leadership and to feel that the founder is sacrificing the hire's 
benefit (value creation) for the founder's own benefit (personal control). Had some of Segway's top 
hires probed their founder-CEO more deeply, they might have avoided being caught in a situation 
where Dean demanded control of all important decisions, even in areas where he lacked business 
expertise that others possessed. 

Likewise, as described in Chapter 10, potential investors have to diagnose whether the founder of 
a startup in which they might invest will be aligned with their own wealth motivations. Down the 
road, if the founder's interests diverge from those of the investors — for example, if the investors 
believe a different CEO could do a better job building the value of the startup — will the founder be 
more on board with the change because it would be financially beneficial for everyone, or will the 
founder fight to retain control, even if it brings down the value of the startup — or possibly brings 
down the startup itself?* 

CAN FOUNDERS CHANGE? 

Evan Williams found it very hard to change his core motivation. Although he seemed to have changed 
from a control-motivated founder at Blogger to a wealth-motivated founder at Odeo, he came to regret 
his loss of control there and began to reverse his course, making choices designed to win back confrol 
of his startup. Once he realized that the problem with Odeo was his inability to make his own 
decisions, due to the powerful influence of his investors, he made the nearly unprecedented decision 
to use his own money to buy out his VCs — that is, to buy back the control he had mistakenly 
surrendered. Not that surrendering control is inherently a mistake, but it was a mistake for Evan 
because it violated his core motivation as a founder. 

As described in Chapter 2, the Career Leader data on entrepreneurial motivations suggest that 
those motivations may be very stable throughout a founder's life. However, people differ in the 
relative strengths of their wealth and control motivations. For some founders, such as Evan and Brian 
Scudamore, one is near or at the top of the list and the other much lower down; for others, the two 
motivations have nearly equal force. The former type would find it much harder than the latter to 
reverse the ordering of their motivations. 



At the same time, the ups and downs of startup life can indeed shape the founder's motivations. On 
the down side of the entrepreneurial roller coaster, even the most control-oriented founder — burned 
out from months or even years of struggle — may be more willing to give up the reins and settle for 
wealth rather than control. James Milmo of Lynx Solutions said that he and his partner, who for 
months had fought each other for control of Lynx, were finally attracted to a buyout offer because "we 
were tired and we needed a break." Although the Ml buyout offer on the table was much less 
lucrative than other acquisition offers, James said that being able to walk away from Lynx was "very 
appealing." On the up side of the ride, a founder who fulfills one objective may change his or her 
enphasis to achieving the other. For instance, a control-oriented founder-CEO who has become rich 
may decide that further financial gain is no longer worth the hard work of being CEO and be willing 
to step aside. At Transcentive, founder-CEO Mike Brody experienced both of these effects. Mike was 
able to step away from the company he had controlled for more than 14 years when Les, the new 
CEO, negotiated a partial founder buyout that allowed Mike to retire as a rich man. Also, after so 
many years at the helm of his company, guiding it from a startup to a more mature conpany, Mike was 
ready for a change and, according to Les, "tired." 

As founders gain knowledge and confidence in their skills, they may also seek to gain more control 
of decisions. For instance, when Tim Westergren first founded Savage Beast (later renamed 
Pandora), he had no prior business experience, having worked as a nanny, musician, and composer. 
He believed that his partner Jon, who had prior startup experience, was far more qualified to lead. 
Over the years, however, Tim's knowledge, skills, and confidence grew. He took on more of Jon's 
responsibilities and became the "evangelizer" who was the public face of Savage Beast's innovative 
service. Much to his own surprise, Tim found himself willing and ready to take over as CEO when 
Jon decided to leave. 

The nature of a new opportunity may also cause the founder to shift his or her priorities. Evan 
Williams realized that podcasting would be a huge market and that other, more advanced companies, 
such as Apple and Yahoo, were sure to enter it with their own technology. Odeo had the early-mover 
advantage, having made a splash at a popular industry conference, and Evan felt a need to press his 
advantage by being first to market — a decision more typical of a wealth-motivated founder than a 
control-motivated founder. In addition, the relative ease with which Evan was able to raise VC 
fimding for this hot technology placed him firmly on the path of fast, aggressive growth. 

As the first-time founder sees the challenges that come with the later stages of startup evolution, he 
or she may also conclude that "I'm a startup person" who loves to get a startup off the ground and then 
move on to the next startup, rather than hanging in for what can seem like the long slog of the later 
stages of consolidation and formalization. 

Founders who go on to found more startups can give us a window into how the learning process 
during a founder's first startup may cause a change in motivations that then shapes future efforts. As 
we have seen, first-time founders are often ignorant of their own capabilities, the outcomes of their 
early decisions, and the dramatic changes in each stage of startup growth. As they learn from their 
initial e?q)eriences, they go from being ad-hoc and naive to systematic and informed. While building 
Wily Technology, Lew Cime came to understand, for exanple, that startups required very different 
skills than more mature conpanies and that a successful startup CEO would not necessarily be the 
best person to lead the conpany as it grows. In Lew's words, he realized that he was more of a 
"speedboat captain" than "captain of an oil tanker." 

While Evan's case shows a founder who traveled 360 degrees from King to Rich and back again, 
Frank Addante experienced a gradual 180-degree shift in priorities over his five startups. As a young 



and inexperienced entrepreneur, Frank was happy to cede control of and equity in his first two 
startups to his more veteran partners. In his third startup, Frank began to actively prepare to become a 
CEO. Finally, in his fourth and fifth startups, he gained control, setting out from the beginning to be 
"The Guy" who was "going to lead the army and go after that market and conquer it" instead of being 
a lesser member of the team 

OTHER CHOICES THAT AFFECT RICH-VERSUS-KEVG OUTCOMES 

Figure 11.1 focuses on the founder's choices regarding other key participants in the startup. But 
wealth-versus-control outcomes are also affected by other choices, some of which are made even 
before founding. These choices include: 

Preferred Rate of Startup Growth 

Founders who want to keep control should prefer slower growth; at the least, the startup should be 
growing no more quickly than the founder's ability to learn the new skills required for each new stage 
of growth. Founders who lack confidence may also prefer slower growth; for some founders, "self- 
doubt about their management capabilities led them to avoid rapid expansion. "^^ (In turn, the 
founder's growth goals should affect the type of financing used by the founder.) In contrast, founders 
who want to maximize startup value should be open to faster growth if that's where the best 
opportunity lies. In some industries and in some stages of the business cycle — for exanple, in 
industries without drastic conpetitive pressures and during down parts of the business cycle — 
founders often have more control over the startup's growth rate. In industries with intense competitive 
pressure or during boom times, on the other hand, the founder's only choice may be "grow fast or 
die." 

Evan Williams was under incredible pressure to develop Odeo's podcasting tool quickly, 
particularly after Apple engineers showed him the prototype of their new rival product, called 
"iTunes." Evan explained why he decided to raise VC capital so early in the startup's life, having 
shunned it altogether in his prior startup: "There was definitely pressure to . . . grow fast. . . . Apple 
came on, and then Yahoo launched a product within the same year, and every major media company 
announced something in podcasting and many of them were calling us, wanting to do deals. There was 
a feeling that this was all moving very, very quickly, and we needed to be ahead of it." 

Capital Intensity 

Similarly, founders who want to keep control should strive to begin startups that have low capital- 
intensity or that require few or no resources beyond those the founder already controls. Founders who 
want to maximize startup value, on the other hand, should be open to pursuing ideas that require high 
capital-intensity and to attracting the necessary resources. 

Boundaries of the Firm 

Capital intensity can be affected by which activities the founders decide to include within the 
boundaries of the firm For instance, a startup could decide either to carry out all core tasks internally 



or else to outsource some of them/" At Ockham, Jim Triandiflou relied on outsourced software 
programmers to develop his sales-ft)rce-optimization product, saving him the trouble of adding a 
technical cofounder and deferring the need to raise capital. Dean Kamen, however, reftised to allow 
outsiders to become involved in Segway's production, even though this decision forced him to raise 
considerable capital and hire several auto industry and production experts to build a complex 
production competency in-house. Franchising is another way to expand the traditional boundaries of 
the firm Franchising can afford founders tangible benefits, such as enabling rapid growth in far-flung 
markets and sharing the risks and costs of expansion with highly motivated owner-operators.^^ 
Franchising allowed Brian Scudamore to build a highly valuable rubbish-removal enpire without 
raising one penny of outside capital or sharing equity with a single cofounder. 

INCREASING THE CHANCES OF ACHIEVING THE "ENTREPRENEURIAL IDEAL" 

Future research could benefit from a focus on what distinguishes those few first-time founders who 
achieve the Holy Grail of Rich-and-King status. We could very well find that the major distinction is 
luck; for example, a rare window of opportunity such as IBM introducing its PC but needing an 
operating system for it and tapping startup Microsoft. However, outside of hoping luck will break 
their way or waiting for a once-in-a-lifetime idea to strike, what can founders do to increase the 
chances they will achieve both Rich and King? 

Increasing Pre-founding Resource Endowments 

Founders who begin their startups with more resources — that is, who made pre-founding career 
choices that enabled them to develop more of the human, social, and financial capitals needed to 
pursue the opportunity — will have a higher probability of achieving the entrepreneurial ideal. One 
such founder is Barry Nails, whom we've gotten to know throughout this book and who will help us 
see how such founders prepare themselves to launch startups: 

Human Capital 

Founders who have accumulated the skills and knowledge needed to start and build their startups 
should be able to go without outside resources — and avoid the costs they impose — for a longer 
period of time. When they do eventually tty to access those resources, they should be able to do so on 
more attractive terms, maintaining more control and giving up less ownership. Specifically, potential 
founders should work long enough to accumulate the managerial and functional skills required to be a 
founder-CEO throughout the stages of startup evolution, including the deep sales skills that become 
more critical as a startup matures. Potential founders should also try to work beforehand in the 
industry in which they are most likely to found. 

After working for a decade at GTE, Barry Nails attempted to start a small generalist consultancy, 
taking on whatever local projects he could find. It ended up being a very rocky experience, and he 
shut it down after a couple of years, returning to GTE. A decade later, when he decided to step back 
into the entrepreneurial waters, he was more prepared. First, he chose to found in an industry — 
telecom — that he knew extremely well. Second, he had gained additional managerial experiences that 
he believed would be attractive to investors: "I knew the technology. I had been an inttapreneur and 



had built a division from $30 million to $1 billion, so I had a real track record for [investors] to 
smell and validate." After GTE, Barry worked at two small startups, again gaining valuable 
experience by learning the pace and working environment at small entrepreneurial firms. 

Social Capital 

Social capital plays a central role in attracting other resource providers. At GTE, Barry Nails had 
gained contacts with many telecom industry players who were potential customers, employees, and 
investors — contacts he leveraged to obtain financial and human capital. More broadly, founders can 
leverage their social capital to enhance both the startup's potential for financial gain and their own 
control of decision making.^^ Entrepreneurs with more social capital earn higher returns from the 
opportunities they pursue^^ and increase the likelihood that their startups will go public. On the 
control side, greater social capital should increase an entrepreneur's confrol over his or her startup. 
Prior ties reduce information asymmetries and investor risk,^^ thus reducing the degree of control 
investors will feel they need in order to protect their investments. Resource providers with prior ties 
to an entrepreneur are usually willing to incur greater risk and to be less likely to take actions 
detrimental to the entrepreneur.^^ Board allegiance, which often derives from prior ties, can forestall 
a CEO's dismissal,-^"^ at least in part because people with prior ties enjoy a deeper level of trust.^^ 

Financial Capital 

Potential founders should be avid savers, adopting a low "personal burn rate" even when they are 
receiving comfortable corporate salaries so they can accumulate as much seed and post-seed capital 
as possible. Such founders should enjoy greater bargaining power when they negotiate with resource 
providers.-^^ An entrepreneur's initial investment of financial capital can help eliminate liquidity 
constraints-^^ and might affect the startup's ultimate level of success. Barry Nails had negotiated a 
six-month severance from his last company, giving himself that long to start a new firm "I gave 
myself six months to put together the complete plan, and raise money from investors," he recalled. "If 
I couldn't do it in that six-month window, I'd go get a job. We wouldn't have starved, wouldn't be on 
the street." Moreover, the positive signal sent by entrepreneurs who invest their own resources 
should improve access to outside resources. Combined with a solid resource base, such positive 
signals should help attract better cofounders and hires. 

Anticipating Trouble 

We have seen how founders' early decisions can drastically hinder their efforts to build value and to 
retain control of the startup. For instance, "playing with fire" by founding with friends or family can 
cause major damage to the startup, the founding team, and the founder's social relationships outside 
the startup. Splitting equity in a static fashion can cause drop-out founders to make off with equity 
desperately needed to replace them or can enable them to hold up much-needed rounds of financing. 
Taking capital from a VC firm that tends to change CEOs at the first hint that the founder-CEO might 
not be able to handle every aspect his or her job can lead to early founder-CEO succession fireworks. 
Founders who understand the long-term risks introduced by these common early decisions can 



anticipate those problems and take action to avoid them. In particular, founders who play with fire by 
cofounding with those close to them (hoping to gain the benefits of doing so) can proactively erect 
strong firewalls to protect themselves and the startup should those personal relationships blow up. 
Founders splitting equity can include terms for buying out cofounders if irreconcilable differences 
arise within the team or can structure dynamic equity splits that adjust to fiiture changes. Before taking 
capital from potential investors, founders can perform their own due diligence on those investors to 
learn from past founders whether, for example, the investor tries to "coach the founder-CEO to 
success" when the startup hits a rough patch or tends to pull the trigger and replace the founder as 
CEO.* Founders who take steps in advance to recover from such problems are more likely to create 
more value, retain more ownership, and/or maintain control, thus increasing the chances of 
maximizing wealth while maintaining control. 

Getting Closer Each Time: Serial Entrepreneurs 

We have focused on first-time founders. Serial entrepreneurs — founders of multiple startups in 
succession — often did not achieve Rich-and-King status in their early startups, but eventually 
accumulated knowledge and resources to help them get closer to achieving the entrepreneurial ideal 
in subsequent startups. For instance, my data show that serial entrepreneurs receive larger equity 
stakes within their founding teams (increasing how wealthy they can become from the startup) and 
remain CEO for longer (increasing their reigns of control) than first-time founders do. As we saw in 
Chapter 9, "Investor Dilemmas," 1 8% of founders take capital from investors who had invested in the 
founder's prior startup. By leveraging such preexisting ties, serial entrepreneurs should also be able 
to negotiate investment terms that leave them more in control and/or owning a larger percentage. 
Other recent research also suggests that serial entrepreneurs have a higher probability of startup 
success.^^ 

Although power and wealth are often at odds in a founder's first startup, his or her increased 
wealth after a successfiil startup may result in "delayed power" in a subsequent startup. Lew Cirne, 
for instance, after handing the reins of Wily Technology to Richard Williams, was sidelined within 
his own startup but a couple years later ended up a rich man when Computer Associates bought Wily 
for hundreds of millions of dollars. In his second startup. New Relic, Lew consciously kept tighter 
control over the early decisions by solo founding, self-ftmding until he had enough leverage to get the 
benefits of VC money without the risk of losing control, and careftilly choosing each person whom he 
involved as a hire or board member. 

FeedBurner's "serial founding team" was able to go from near failure in its first startup to 
achieving better results in each subsequent startup until it achieved the entrepreneurial ideal in its 
fourth startup. Along the way, the team learned lessons at every level at which our founding dilemmas 
occur. Team members ironed out early role problems and learned how best to split the equity. With 
hires, they learned not to hire specialists too soon, how to strike the right balance between flexibility 
and depth, and how different ftinctions develop at different rates and have to be compensated 
differently. They learned how to gain bargaining leverage with investors, which terms were most 
important to negotiate, and how to build, manage, and maintain control over a board. As a result, 
these founders retained control throughout the evolution of FeedBurner while leading the startup to a 
lucrative exit. 



IMPLICATIONS FOR "UNSOLVED PUZZLES" 

In Chapter 1, we previewed several unsolved puzzles concerning founders. We are now in a position 
to shed light on each of them 

Is the Missing "Private Equity Premium " Really a "Puzzle "? 

Microeconomists have wondered why entrepreneurs, despite the higher risks they take and their 
presumed financial motivations, do not tend to make more money than they could in paid employment 
— probably even less on a risk-adjusted basis. The wealth-versus-control trade-off provides a 
straightforward and empirically supported explanation for this lack of an entrepreneurship premium: 
Although some founders are motivated to maximize their financial gains, many other founders are 
motivated by control, passion for the idea, and other motivations and therefore make decisions that 
sacrifice financial gain in order to maintain control or to maximize other nonfinancial benefits. 
Research on entrepreneurs should become less puzzling when researchers take into account the 
different kinds of success different entrepreneurs are motivated to seek. 

Founder Power? 

In academic sociological research, an executive's status as a founder has been used as a proxy for his 
or her greater power within the startup. However, my research shows that founder status can 
indicate exactly the opposite. Founders' attachment to their startups causes them to suffer financially 
and to lose control of their startups (as described in Chapters 6 and 10, respectively). This is one 
exanple of how the entrepreneurial arena can help us shed light on power-elite theory, which 
assumes that the corporate power and economic wealth of top executives reinforce each other, with 
corporate positions serving as the source of wealth while "money provides power."^^ In contrast, the 
research described throughout this book highlights recurring conflicts between wealth and power, 
examining the circumstances in which they are in tension. 

We have also seen how early sources of a founder's power can later become liabilities.* 
Founders' early successes ofl;en depend upon their optimism, persistence, and passion. However, if 
they are not careftil, these very traits can skew their decisions and lead them to make mistakes at 
every stage of startup evolution. For instance, optimism may mislead potential founders into making 
the leap before they are ready (see Chapter 2), underestimating their need for cofounders (see Chapter 
3), discounting the dangers of founding with fi-iends and relatives (see Chapter 4), misjudging their 
ability to be effective CEOs through multiple stages of startup evolution (see Chapters 5 and 10), 
fighting for larger equity stakes than they objectively deserve (see Chapter 6), failing to fill holes 
with appropriate hires or even to recognize those holes (see Chapter 8), or agreeing to investment 
terms (e.g., liquidation preferences) that they think will never apply to them (see Chapter 9). Any of 
these and other missteps can weaken the founder's power, harm the value of his or her startup, and 
even imperil his or her ability to remain involved or for the startup to survive. 

Stewards Rather Than Agents? 

The predominant lens in microeconomics studies is agency theory, which posits that problems are 



caused by misalignment between the individual "agent" and the organization (or its owners). The 
theory proposes that organizations can reduce agency problems by means such as incentive alignment 
and monitoring. Stewardship theory, on the other hand, posits that, in certain contexts, executives' 
interests will be aligned with company interests and be more intrinsically motivated than predicted by 
agency theory. "^"^ The findings in this book suggest that, in many cases, stewardship theory can better 
represent the relationship between founders and their startups and more accurately predict founder 
behavior than agency theory can. Even so, my research exposes a dark side of stewardship theory, for 
the close attachment between founder- stewards and their startups causes problems for founders 
regarding both wealth (resulting in a greater founder compensation discount) and control (causing 
founder-CEO succession to be even more disruptive to the founder and the startup). Founders' 
stewardship inclinations can negate their power and even become destructive. 

BOUNDARY CONDITIONS AND OPPORTUNITIES FOR FUTURE RESEARCH 

The data and case studies at the core of this book come from for-profit American startups in the two 
largest industries for high-potential startups — technology and life sciences. The dilemmas examined 
throughout the book apply specifically to the founders of such startups. It is still unknown whether and 
to what extent they are specific to the United States. We have also focused on subsets of the possible 
motivations, types of founders, and founding dilemmas. These limitations and boundary conditions 
must be kept in mind regarding the results and patterns described in this book, but also suggest 
avenues for fiiture research. 

Testing Boundary Conditions 

The two major boundary conditions are the types of organizations examined here and the country from 
which the data were collected. 

Specific to High-Potential Startups? 

Both the field cases and the quantitative data presented in this book come from high-potential 
startups.* In such startups, founding dilemmas are often stark, for the founders are usually very 
resource-constrained, rarely have all of the skills and connections needed for the startup to reach its 
full potential, and tend to face significant risk of losing decision-making confrol. Some studies of 
small businesses have begun to suggest that, even in low-potential startups, founders may sacrifice 
financial gains for nonfinancial benefits. However, these researchers have not directly examined if 
and to what extent small-business enfrepreneurs give up confrol of their organizations and, if they do 
give up confrol, whether they gain financially from doing so. Answering these questions would help 
establish whether the dilemmas faced by those small-business founders differ from the dilemmas 
examined here. 

Nonprofit organizations and family businesses also offer good opportunities to examine the tension 
between wealth and control. Nonprofits face the core enfrepreneurial challenge of pursuing 
opportunity while having to gain confrol of necessary resources,^^ suggesting that the founders of 
nonprofits may face dilemmas similar to those examined in this book. As nonprofits grow, they also 
face distinct changes in the skills needed to lead the organization, posing major challenges for the 



founders' abilities to continue leading them. However, nonprofits and their founders may face even 
more extreme challenges because the founders may become even more central to their organizations 
than other founders do, more attached to them, and more identified with their missions, making it even 
harder to transition to a new generation of leadership. Also, many nonprofit founders are motivated 
by "impact" and the chance to improve things for the community or the world. Maximizing that impact 
calls for attracting the necessary resources, but that, in turn, may imperil the founder's control of 

decision making. Do nonprofit founders face a trade-off between impact and control that parallels 
the wealth- versus-control trade-off we have been examining, or does that pair of motivations result in 
different trade-offs and different dilemmas for nonprofit founders? 

In the family business realm, does the motivation to maintain control of the business so that 
descendants can continue to run it introduce trade-offs that parallel the trade-offs examined here, or 
does it introduce different dilemmas? If multiple members of the family have a say in the business, but 
differ in their motivations, how do they resolve those differences? Not only would a wealth- versus- 
control lens help us deepen our understanding of nonprofits and family businesses, but that deeper 
understanding should in turn enrich our understanding of the core dilemmas encountered in high- 
potential startups. 

Specific to the United States? 

Almost all of the cases and data in this book come from the United States. We do not know yet 
whether the dilemmas examined here are specific to the American culture, legal regime, and 
regulatory structure or whether they apply more universally. Culture certainly affects the propensity 
for founders to found alone, to do so with family members, to split roles in an egalitarian fashion, to 
split equity equally, to hire inexperienced people, and to avoid (or not have access to) professional 
investors. Whether the differences mean that the actual dilemmas differ, or just that the different 
choices within each dilemma occur with a different frequency, remains to be established. 

At the same time, data collected by the Global Entrepreneurship Monitor (GEM) project, a multi- 
institution effort to understand entrepreneurship around the globe, suggests that most countries have 
similar mixes of Rich versus King founders. GEM surveys entrepreneurs about their core motivations, 
categorizing them as being motivated by "independence" (i.e., control) or by the desire to "increase 
income" (i.e., wealth). Across the 42 countries in GEM's report, including 23 high-income 
countries and 19 middle- and low-income countries, almost every country had a relatively even split 
between the two types of motivation. In every country, at least 35% of entrepreneurs were control- 
motivated and, in all but two, at least 30% were wealth-motivated.^^ The split in the United States — 
about 57%) control motivated and 43 %o wealth motivated — ^was similar enough to the splits in other 
countries to suggest that the U.S. -based results in this book might extend to many of those other 
countries. But we still need solid evidence that this is frue. 

Opportunities for Future Research 

Throughout The Founder's Dilemmas, I have presented the results of a wide variety of analyses of 
my quantitative dataset. These results provide an initial look at how the patterns described here affect 
such outcomes as the propensity to found alone, founding-team stability, startup valuation, and the 
ability to remain CEO of the startup. Future multivariate research could (a) test other important 



outcomes and (b) test other propositions about how these outcomes are affected by the variables and 
contingencies described in this book. Other opportunities for high-potential research include 
broadening the picture to include other entrepreneurial motivations, the experiences of serial 
entrepreneurs, other participants in the building of startups, other dilemmas faced by founders, and the 
relative risks and returns of founder exits. Because any honest model of a conplex human 
phenomenon has to acknowledge the many unknowns, these opportunities for future research are 
integrated into the cohesive model of founder decisions presented in Appendix 1 1-2. 

Beyond Wealth and Control: The Effects of Other Motivations 

Although wealth and control motivations are the most common for entrepreneurs, they can coexist 
with others. For instance, a founder who is equally motivated by wealth and control motivations, and 
thus lacks clarity about what to decide regarding a specific dilemma, may use a secondary motivator, 
such as intellectual challenge, as a tiebreaker, which might lead to different decisions than those made 
by a founder who uses yet another motivation — say, altruism — to break the same tie. Also, the 
framework in this book applies to the majority of founders, who are indeed wealth- or control- 
motivated, but not to founders with less common motivations.* It thus cannot provide solid guidance 
to those founders about what decisions to make when facing the dilemmas described in this book. 
Developing such guidance requires both the theoretical groundwork to understand the linkages 
between their motivations and the decisions they should make, and the empirical testing to validate 
those relationships. 

Beyond First-Time Founders: Serial Entrepreneurs hip 

My consistent focus has been on first-time founders, occasionally contrasting them with serial 
entrepreneurs such as Dick Costolo, Evan Williams, Lew Cirne, and Frank Addante. Academia has 

begun taking a more concentrated approach to studying serial entrepreneurs,^^ and, for each of the 
dilemmas examined here, there is a fertile opportunity to research the systematic differences between 
first-timers and serial entrepreneurs. In what ways does the decision making of serial entrepreneurs 
evolve? How do their motivations evolve, and is that affected by how successful they were in their 
initial startups? Do any of these changes create underappreciated pitfalls for the serial entrepreneur? 

Beyond Founders, Hires, and Investors: Other Potential Participants 

This book has focused on the most central participants whom founders might involve in their startups: 
cofounders, hires, and investors. However, other participants might also play important roles as 
alternative resource providers. For instance, rather than gaining full control over some resources, 
founders may create corporate partnerships with companies that have complementary resources or 
companies to which they can outsource important tasks. These founders will thus face "whether to 
partner" dilemmas. 

Such partnerships may involve different wealth-versus-control trade-offs than those involved in 
attracting cofounders, hires, and investors. For instance, crafting a corporate partnership may cause a 
founder to give up control of some decision-making power (because the startup will no longer be 
doing everything in-house) and of some financial gains (which will now have to be shared with the 



partner), but might increase the founder's chances of retaining a moderate amount of control (because 
he or she didn't have to take as much capital from outside investors) and of securing at least some 
financial gains from his or her hard work (by staying focused on the core of the business, remaining 
nimble, getting to market and scaling the business more quickly, or keeping costs down). In this sense, 
corporate partnerships may be a way to achieve "75% Rich, 75% King."^^ Such participant issues 
can become "boundary of the firm" decisions that can also have important implications for how much 
control the founder can keep over decision making and how valuable a startup he or she can build. 

Where to Found, Exits, and Other Important Dilemmas 

A dilemma that is not addressed in this book is one of the earliest decisions a founder faces: where to 
found the startup. Locating in hubs can help founders attract the best cofounders and hires (who 
might already be living there and not want to move elsewhere) and the best investors (who like to 
invest in clusters of nearby startups). However, if such a hub is not the best place for the founder or 
his or her family, he or she may decide to locate where the resources are not as available and thus 
limit the potential value the startup can achieve. (This is one way in which the personal, career, and 
market factors from Chapter 2 can affect eventual outcomes.) Barry Nails returned to North Texas 
from California because of the resources available there for his autistic son, even though California 
might have been a better place to find the resources he would need for his startup. Genevieve Thiers 
moved Sittercity from Boston, where she had contacts and employees, to Chicago, both to pursue her 
avocation in opera at a master's program at Northwestern and because her fiance was born there and 
wanted to move back home. The FeedBurner team was adamant about staying in Chicago for their 
families' stability, even to the point of negotiating specific buyout terms that would ensure they would 
not have to move. 

A final dilemma that deserves attention is whether to sell the startup to an acquirer or to go public. 
Appendix 11-3 presents an initial field-based exploration of these exit decisions, but much more 
research is required to establish the patterns and best practices. How do wealth- versus-control trade- 
offs affect such "exit dilemmas"? How are the founders' decisions affected by their tendencies 
toward overconfidence, optimism, and attachment? How do nonfounders, including professional 
CEOs and investors, affect the exit decision, and how often do their interests diverge from those of 
the founders? What are the most effective mechanisms for resolving such divergences (e.g., founder 
carve-outs or earn-outs)?^^ 

Other Outcomes: Realized Exits and Risk Differences 

Quantitative examination of founders' exit dilemmas would have a fiarther benefit. Because the data in 
this book come from operating startups, we can compute only the current value of the founders' equity 
stakes — their "paper gains" — ^but not their "realized" values — the wealth eventually received from 
the startup. A dataset that included startups that had exited and returned actual money to their 
shareholders (a much smaller dataset than the one used here) would help us to assess how closely 
startups' interim paper gains approximate their eventual realized gains. 

In addition, although we have deemed some choices riskier than others (e.g., founding with friends 
or family members and doing quick-handshake equity splits) and have some information about 
outcomes such as the breakup of a founding team, we don't have the final outcomes that would give us 



the last word, so to speak, on which decisions are safer than others. For instance, how often is a 
control-oriented ft)under able to sell a startup, of which he or she owns 100%, for $5 million, and 
how often is a wealth-oriented founder able to sell a startup, of which he or she owns 5%, for $100 
million? How much longer do control-motivated founders take to exit compared to wealth-motivated 
founders? Is one exit a longer shot than the other? If so, how should differences in risk profile affect 
the founder's early decisions?* 

CLOSING REMARKS 

There is a great deal of mystique around founders, some of it well deserved but some of it also 
dangerously unexamined. Many founders and would-be founders see themselves — and are seen by 
others — as having a calling. This is true and very inspiring, but founders — and those who join their 
startups — need a clearer picture than that. Different motivations call for very different decisions. 
Many founders have learned the hard way how their own choices undermined their motivations. As 
the Roman playwright and philosopher Seneca wisely observed, "If one does not know to which port 
one is sailing, no wind is favorable." 

But knowing one's motivation is only one piece of the puzzle. Founders often embark on their 
entrepreneurial journeys feeling like Lewis and Clark as they set out across the continent, having a 
rough idea of where they want to go, but seeing no clear road ahead nor ftilly anticipating the dangers 
they might encounter. Once on their way, they often fail to realize when they are at a crucial fork in 
the road they are on, about to make a decision that will have inportant consequences for the journey. 
Perhaps they see a path to the right but not the path to the left. Nor do they see that, just around the 
bend, the path to the right heads straight over a cliff. Neither instinct nor luck can consistently guide 
their way, mile after mile, as reliably as a clear understanding beforehand of the many forks in the 
road, the options at each fork, and the implications of each of those options. 

I have called these forks "founding dilemmas," which sounds more like a warning than an 
invitation. But they are not unsolvable. Scientific study and the wisdom of the experienced can each 
help. I have gathered as much as I can of both into this book. As an academic, I am interested in both 
success and failure, but to those who use this book, I wish only success on the path ahead. 



* See Chapters 1 and 2 for empirical evidence of the preponderance of these two motivations. 

* Likewise, in power-elite theory, the "corporate rich" — powerful and wealthy top executives — are also a very small, elite subset. 

* Throughout this book, we have looked at how overconFidence can skew founding decisions. Overconfidence can also mislead 
founders about the probability of achieving the entrepreneurial ideal, leading them to attend to only those exceptional founders who grew 
valuable startups while maintaining control throughout, or else to assume that their own chances are higher than the typical founder's. 
The choices they make because of this misperception may increase the chances of their failing, rather than enabling them to achieve 
their top priority, whether it be maintaining control or maximizing wealth. 

* From a revealed-preferences perspective, in which we examine people's actions to learn about their preferences, solo founding is 
one of the clearest indicators of a founder's control orientation. However, it does not enable us to separate solo founders who intended to 
(and did) found alone from those who tried to find cofounders but failed. (Given the "noise" introduced by this, the analyses described 
here should be conservative, underreporting the statistical significance of the independent variables.) As described below, future research 
should examine fUrther indicators of control motivations, possibly including self-reported measures of motivation or survey questions 
similar to the CareerLeader survey questions detailed in Chapter 2. 

* Of course, it may be a rare founder who is able to smoothly move back and forth between such opposing actions! 

* Likewise, in Life of Caesar, Plutarch quotes Julius Caesar's clarity about his earfy need to accept a smaller kingdom in order to 
have control of it: "For my part, I had rather be the first man among these [lesser] fellows, than the second man in Rome." 



* See Chapter 10 for a discussion of the need for potential successor-CEOs to diagnose the founder. 

* From the founder's perspective, we have argued that wealth and control motivations are equally valid as long as the founder's 
decisions are consistent with those motivations and with the overall context in which the startup is operating. However, as a society, 
should we care what motivates our entrepreneurs as long as they continue to be motivated to found startups? Very possibly. There is a 
tension between societal goals and some of the personal goals on which we have focused. A wealth-motivated founder makes decisions 
that maximize value creation, which may be aligned with societal interests. A control-motivated founder, however, sacrifices value 
creation in favor of maximizing the chances of getting what he or she most prizes. A founder who has an idea with world-changing 
potential could build a startup that employs hundreds or thousands of people and delivers tremendous value to customers, but if that 
founder is control-motivated, he or she might make decisions that impede that value creation. 

* In recent years, websites such as Adeo Ressi's TheFunded.com have made it much easier for founders to perform such due 
diligence on their potential investors, further increasing their abilities to proactrvely prevent such problems. 

* Morgan McCaU's research on executive derailment — when successful managers who were expected to continue being successful 
fail instead — highlights four common dynamics that have parallels to some of the dynamics described here. McCall's four are that the 
managers' prior strengths become weaknesses in a new situation, their existing flaws do not become salient until they are in a new 
situation, their success leads to arrogance, and they experience bad luck. For more details, see McCall (1998). 

* As defined in Chapter 1, high-potential startups are startups, often technology- or science-based, that have the potential to become 
large and valuable, even though their founders may subsequently make decisions that limit their growth. See Chapter 1 for a brief 
discussion of the differences between high-potential startups and small businesses. 

* As shown by the CareerLeader data in Chapter 2, wealth and control motivations dominate the list of entrepreneurial motivations, 
but altruism, variety, and intellectual challenge are also top-four motivations for some age cohorts. 

* In Hall et al.'s (2010) study of VC-backed startups, 75% of the startups returned $0 to the founders. How does that compare to the 
return to King founders, and how do the risks of losing the CEO position differ for each type of founder? 



APPENDIX 11-1— A THOUGHT EXPERIMENT: WOULD A HYBRID PATH BE BETTER? 



Let's engage in a brief thought experiment regarding founder decision making. In this sinple world, a 
founder has decisions to make about each of three people to involve in the startup — one decision 
about cofounders, one about hires, and one about investors — and has two choices within each 
decision — a control-oriented choice, which maintains control at the risk of not attracting the person, 
and a wealth-oriented choice, which attracts the person but heightens the risk of losing control. Let's 
assume that each control-oriented choice will result in an 80% chance of maintaining control and a 
20% chance of building the full value possible, and that each wealth-oriented choice will result in a 
20% chance of maintaining control and an 80% chance of building the full value possible. 

If the founder consistently makes the control-oriented choices, it will lead to the following 
probabilities of each result: 

• Maintain control: 80% x 80% x 80% = 51% probability 

• Build fullest value: 20% x 20% x 20% = 1% probability 

In contrast, if the founder consistently makes the wealth-oriented choices, it will lead to the 
following probabilities of each result: 

• Maintain control: 20% x 20% x 20% = 1% probability 

• Build fullest value: 80% x 80% x 80% = 51% probability 

Founders who make consistent choices across all three levels will thus be approximately 5 1 times 
more likely to achieve one result than the other. According to this model, these founders will also 
have a slightly better than 50/50 shot at reaching one of the desired outcomes, rather than failing to 
reach either outcome, but also a less than 1% chance of achieving both outcomes. 

If the founder is less consistent about his or her decisions, the probabilities change markedly. If a 
founder chooses control for two of the decisions and wealth for the third (though not necessarily in 
that order), it will lead to the following probabilities of each result: 

• Maintain control: 80% x 80% x 20% = 13% probability 

• Build fullest value: 20% x 20% x 80% = 3% probability 

The chance of maintaining control has dropped by a factor of four, while the chance of building the 
full value of the startup has risen to only 3%. The chance of complete failure — not achieving either 
outcome — has risen dramatically, to better than an 80% chance. Likewise, for one control-oriented 
choice and two wealth-oriented choices, there will be a 3Vo chance of maintaining control and a 13%) 
chance of building the startup's full value. 

A hybrid strategy may indeed provide one benefit: Founders who aren't sure which outcome they 
prefer can "hedge" their bets by mixing some wealth-oriented choices with some control-oriented 
ones. By making consistent choices, founders have only a 1% chance of achieving the less likely 
outcome, but by making mixed choices, they can increase that probability to 3%. However, the cost of 
this hedging is dramatic, causing the probability of achieving the more likely outcome to drop 
dramatically and increasing dramatically the chance of achieving neither of the desirable outcomes. 



For founders who aren't sure which outcome is more valuable to them, such a hedge might make 
sense; however, if those founders could figure out their core motivation before making those 
decisions, they could dramatically increase their probability of achieving the more attractive 
outcome. 

Future research could examine enpirically how often founders use consistent versus hybrid 
strategies and begin establishing whether hybrid strategies are the best or the worst of both worlds. 



APPENDIX 11-2— A MODEL OF FOUNDER DECISIONS 



Throughout this book, I have described a wide variety of decisions faced by founders and the factors 
that might affect the outcomes they achieve. These factors range from the founder's motivations and 
behaviors to strategic and environmental influences. Where relevant, I have presented my initial 
research results and provided suggestions for fiiture research. Figure 11.4 integrates all of these 
factors into a cohesive model of founder decisions. 

The figure begins with the founder's motivations, which include the wealth and control motivations 
on which we have focused as well as altruism, variety, intellectual challenge, and others. Those 
motivations, shaped by strategic factors, affect founders' pre-founding career decisions; their 
founding-team, hiring, and investor decisions; and their decisions about where to found, whether and 
how to form corporate partnerships, and whether and how to exit from the startup. The interplay of 
these decisions with the founders' own capabilities and consistency and with their teams' 
interpersonal dynamics affects the outcomes for the individual founders, their teams, and their 
startups. Meanwhile, founder motivations, decisions, and outcomes are all affected by the 
environmental factors described throughout the book. To top it off, this whole tangle of interactions is 
dynamic. Founders learn from the results of their decisions, adjust their goals, and develop fiirther 
skills and contacts that, in turn, affect their motivations and subsequent decisions. 

There's a lot going on — that's one of the most important messages of this book. A less obvious 
message of this book is that we have a lot to learn. Throughout the book, I have presented the results 
of my initial research and of research conducted by others, but I have frequently pointed out where the 
necessary research has yet to be done. As indicated by the asterisks throughout Figure 11.4, 
productive opportunities for fiiture research abound. I invite founders to help remove those asterisks 
by sharing their own insights about the many factors we have seen at play and which they will have 
experienced firsthand. I invite researchers to help remove asterisks with their fiiture investigations 
into those factors. 



Intervening Variables 

■ ConiiScflcvofChoiMttCht. 7, 11) 

• team Dynamics* 

■ CapaMibesand tiiecutionof 
Dec sons* 

• tntrtprenfufialLodi* 



Fourxier Mottvattons 

Wtalttt and Control (Oti. 2-t I) 

• hnanaalGara 

• Autonoiny. Pawer& 
Influence. Manayng People 

Other Motwationt' 

• AMtutun 

• Vanety 

• HtlectualChalenge 

• Be 



Founder Dedsloro 

• C*re*r Oeouons (Ch. 2) 

• Whether When. What to found 
■ Fau)(tng-Team Decisions (Chs. i-f>] 

' Relattonshfts. Roles. Rewards 

• Mnng Decisions (Ch.8| 

■ RelatlonshfB, Roles. Hewanls 

• Investor DeasonslCh. 9) 

• Who, When. Whattenm 

• W«h Whom to Partner* 

• Busness-strategy DecKiofts * 



Outcoints 

forVitfountltr 

• Kmg (remain foundef-CtO — Ch.lO, 
maintan board control — Ch.9| 

• RicM'onpapef'-Ch.ll. ateMt*) 

• Fojnd»f'tSal1tf»et«on with Outcome 
(Moo va ti on -Choce consistency)* 

• Rtskmess of Outcomes* 
forthitwmor ttortup 

' Foundng Team StabAty (Cht. 4, 7) 

• Valuation Growth (Chs. 6. 9} 

• Employment Growth* 



Environmental Factors: 
' CompetMve/nme Pressure* 
' tinflronmen ta l Turbulence* 
' tnvtonmeoial Complexcy* 
' CiAural Influences* 
' Regulatory and Legal Structures* 



* Iruiifulfocusforfutureresearth: dscussed n book but 
not analyzed empincaly. 

• Busness-strategy Deosions ndude where to fotaid. 
0h*rKttntt)es of the startup's core idea andbusness 
mode^ capaal nensitv, and boundar«s of the fircn. 



Figure 11.4. A Model of Founder Decisions 



APPENDIX 11-3— EXIT DILEMMAS: WHICH WAY OUT? 



When launching their startups, many founders dream of the Promised Land of going public, seeing the 
sale of equity to public shareholders as the ultimate exit. However, as we saw in Chapter 9, even 
startups that manage to attract venture capital are almost three times as likely to be acquired as they 
are to go public* 

Our focus has been on the dilemmas faced by founders when deciding whether and how to involve 
new players in their startups. A founding team's decision about whether, when, and how to exit can 
be deeply affected — and constrained — ^by those earlier decisions about bringing in cofounders, hires, 
investors, and professional CEOs. Indeed, some of the long-term wealth and control implications of 
those early founding decisions may become evident only as the startup nears the finish line. For 
instance, early decisions may affect how strong a voice the founders have in the exit decision, the 
extent to which they disagree among themselves about it, and the contribution they are able or willing 
to make after the exit. 

Yet academic research has paid much more attention to the investors and acquirers than to the 
founders, and we lack a detailed picture of the factors that affect founders' decisions to exit. To 
address this gap, I have conducted early-stage field research into the major factors that should be 
considered by the lucky few founders who face — and have a strong voice in — the exit decision. In 
this appendix, I describe the initial results of this field research, beginning with acquisition, the more 
common option, and then describing the initial public offering (IPO) option. 

Should We SeU Our Startup? 

Not only is acquisition a more likely outcome than an IPO, but it is also far more of an "exif ' for the 
startup and for the founders. Once a startup gets to the point where it attracts an offer from a larger 
player, it faces a dilemma that could have dramatic implications for the company and for the founders. 
When deciding whether or not to accept an acquisition offer, founders have to weigh the following 
factors (summarized in Figure 11 .5). 



Reasons Not to Sell ^ 



Reasons to Sell 



• Will lose control of decision making 

• Will lose chance to build further value, 
then sell at a better price 

• May not be able to preserve startup's 
culture 

• May not gain financially if investment 
terms (e.g., liquidation preferences) mean 
founders won't share in exit payments 



Eroding prospects for startup: 

• Adverse market changes or dangerously low cash 

• Founder bumout or rising tensions between founders 

• May soon lose control of exit decision and terms 
Attractive post-acquisition possibilities: 

• Achieve wealth goals 

• Use as stepping stone to become serial entrepreneur 

• Gain resources to build startup 

• Work for attractive parent company 



Figure 11.5. Reasons Not to SeU versus Reasons to Sell 



Reasons Not to SeU 



The founders' core motivations have played a fiindamental role in the decisions we have examined 



throughout this book and they do so in the exit decision as well. Control-motivated founders may not 
want to sell because it means losing a lot of decision-making control. Some fear that being 
swallowed by a large company will spoil the culture they have spent years building. Acquiring 
conpanies also usually want to control strategic decisions and to decide who should remain with the 
startup as it is integrated into the acquirer's core business. Early in the life of Blogger, Evan 
Williams received an acquisition offer of $ 1 million. For Evan, selling Blogger would have been the 
way out of several sticky problems; he was dangerously short on cash, and he would soon need to lay 
off his entire staff. His cofounder and employees pushed strongly for the sale. But Evan couldn't give 
up on his baby and decided not to sell. Even when Google made an offer months later, Evan initially 
rejected it: "After four years of pouring my heart into Blogger, I saw a lot of risk in giving up 

control. "^^ 

Yet even wealth-motivated founders, offered a payday after years of hard work, may not want to 
sell. Some believe they can still build significant value by remaining independent. For example, when 
the acquisition market for Internet companies started to improve in the mid-2000s, FeedBurner started 
to attract acquisition offers, but cofounder-CEO Dick Costolo wasn't ready to cash out: "I wanted to 
stay in because I felt there was a lot more stock value to be built," he explained. In 2008, when 
Facebook made a lucrative offer for Twitter, Evan Williams successfully pushed his investors to 
refiise the offer because of "what we could accomplish by when [and] why there's still so much we 
have lefttodo."^^ 

While Dick's decision not to sell was founded on years of working in technology startups, other 
founders' decisions are driven by their optimism and passion. They may have unduly rosy pictures of 
the startup's potential, causing them to underestimate their upcoming growth challenges, to misread 
the competitive landscape, or to anticipate a better offer that never materializes. 

Just as early decisions can affect a founder-CEO's options in the event of his or her own exit from 
the role (the founder-CEO succession discussed in Chapter 10), they can also affect the founding 
team's exit options. For instance, founders who signed term sheets with liquidation preferences that 
prevent them from getting any proceeds from the sale until the investors get back the capital they 
invested will usually resist mightily any offers below that amount, finding numerous ways to hold up 
the acquisition. At Lynx Solutions, cofounders James and Javier had liquidation terms that would 
have required that their investors be paid back their investments of almost $30 million before the 
founders' and employees' equity stakes could be honored. Lynx's two potential offers, both for a lot 
less than $30 million, would have left the founders without any reward after years of hard work, many 
of those years without pay. "The reason we became entrepreneurs in the first place and had survived 
all of our ordeals was that we never lost sight of the upside," explained James, "and if we did have 
some kind of exit, we didn't want to end up with nothing in the end, didn't want to become victims of 
a $30 million overhang." 

Reasons to Sell 

Major reasons to sell include both eroding prospects for the startup and attractive post- acquisition 
possibilities, but may also include more intangible factors, such as founder burnout and 
inconpatibility. 



Eroding Prospects for the Startup 



The founders may be worried about the startup's future, particularly if cash is running low, as was the 
case at DKA, Dick Costolo's first startup, when the firm's strategic investor decided not to fiind a B- 
round and other potential investors also shied away. Eventually, a sale was the only viable option, 
short of closing the company, and the founder-CEO was able to find an acquirer at the last minute. 
"Our CEO pulled a rabbit out of his hat," explained cofounder Dick Costolo. "Selling became our 
only avenue because of how we had financed the company before." 

Founders may also worry that changes in the industry will threaten the startup's core business. Les 
Trachtman, CEO of Transcentive, a stock-option software company, explained how various forces 
changed the stock-option industry. "The market in 2001 was saturated, legislation had passed that was 
cracking down on stock options, and the market was demanding a global presence. Our customers 
were moving from stock-option plans to overall compensation plans, which meant that we would 
need to fill in holes in our offerings. To do that, we either needed to invest heavily in our software or 
find companies we could buy to fill those holes." When the board rejected his idea to acquire a 
company, Les found a company interested in acquiring Transcentive. 

Within the startup team, years of pressure may be eroding the relationship among the founders or 
within the broader executive team As these relationships suffer, the startup often does, too. Under 
such circumstances, an acquisition offer can feel like a lifeline, a chance for the team to solve its 
internal problems — ^by having one or more players leave or be reassigned — ^while also benefiting 
financially. At Lynx Solutions, which had recently become profitable after several years of losses, the 
burned-out and bickering cofounders decided that selling the company would be the best resolution of 
their irreconcilable differences. 

Some founders choose to exit while they still have the power to drive the acquisition negotiations. 
Waiting until after another round of financing may mean that their equity share or their control of the 
board has dipped below the point where they could negotiate the acquisition terms most important to 
them. For Dick Costolo, the inflection point came when he was contemplating the choice between an 
offer from Google to buy FeedBurner or a C-round term sheet from VCs. Dick chose the buyout over 
the financing, in part because, after the C-round, the investors would own over 50% of the company 
and have a three-seat voting bloc on the board. For Dick it was sell now, when he had full control, or 
wait and perhaps be forced by his investors to accept a deal that would not be as beneficial to the 
founding team. Founders who think ahead about the point at which they are likely to become 
secondary players in the exit decision should be in a better position to control that decision.* 

Attractive Post-acquisition Possibilities 

Rather than running away from an eroding startup, some founders are running toward an attractive 
post-acquisition situation. For wealth-motivated founders, an attractive offer is the goal they set out to 
reach years before — walking away from the startup with a lot of money. Even control-motivated 
founders may be happy to take such an offer if they see it as a stepping stone to their next stage of life. 
They could use their payday to shift gears in life or to become serial entrepreneurs, building a new 
startup over which their increased financial capital will give them more control. For instance, after 
handing over the reins of Wily Technology to Richard Williams (see Chapter 10), Lew Cirne 
remained with Wily as founder-CTO. When Wily was sold to Computer Associates in 2006 for $375 
million. Lew stayed a year, then left to found New Relic, another performance-management software 
conpany. Lew explained that he viewed New Relic as a "do-over" in which he could correct the 
mistakes he had made with Wily. Most importantiy, he built New Relic as a soflware-as-a-service 



(SaaS) or subscription- software business, which gave him lower up-front costs, much higher 
recurring revenue, and a business model that fit his skills. Lew was thus able to self-fiind the company 
during its early stages and to maintain decision-making control longer. 

The post-acquisition possibilities may also be more attractive if the acquisition enables the startup 
to gain valuable resources it would not otherwise have had. Both Evan at Blogger and Dick at 
FeedBurner felt that being acquired by Google would provide the financial capital to fiilfill their 
visions and would put the best and brightest engineers and business minds to work on their projects. 

For founders who have put in years of long intense hours, a steady job at a larger company may 
look attractive. For some, it offers a better balance between work and personal life without having to 
abandon the startup to which they are still attached. If the acquirer respects, rewards, and fosters the 
founders' strengths, they may also be glad to work for that company. For instance, two of 
FeedBurner 's engineering- focused founders were excited about being acquired by Google and 
working for an innovative and fast-growing company. 

How Should We Structure the Sale? 

Acquisitions can be structured in a wide variety of ways, each of which may cause repercussions for 
the founders. Two of the major options are to structure it as a cash-only deal, in which the acquirer 
pays shareholders for their equity stakes, or as an earn-out, in which a significant share of the 
potential payoff for the startup's key executives (including founders) comes in the form of 
performance-dependent payments after the acquisition. For an acquirer planning to replace the 
startup's management with its own people, a cash-only deal often makes sense. However, if the 
acquirer wants to retain some of the startup's people or if it senses that using an earn-out structure 
will help it filter out any startup whose own executives lack confidence in its prospects, an earn-out 
structure is more attractive. It lets founders who are still confident in their startup's prospects benefit 
from the acquisition right away (from the non-earn-out part of the offer) while preserving fiiture 
upside (through the earn-out). Acquirers who can tie a significant portion of the acquisition price to 
fiiture performance are often willing to pay a higher price for the startup. 

At Lynx Solutions, the founders faced competing offers from two wireless companies, Mobilink 
and Spotlight.* Mobilink wanted to close within 30 days and to buy out and replace the cofounders 
immediately, but offered only $20 million in cash. Spotlight offered $25 million in cash and stock 
plus a $10 million earn-out upon Lynx meeting challenging milestones over the subsequent two years. 
The founders would remain working for those two years, focused on achieving those milestones. 

Earn-outs appeal to confident founders and sometimes reward them handsomely, but can also come 
back to bite them At one startup, the earn-out was tied to the founders hitting $150 million in 
revenues, the amount they had projected before the acquisition. But they achieved only $59 million 
and would have done better to take a much lower offer with more of the money up-front. 

Founders who remain with the startup to build more value face a more fundamental risk: They will 
lose a certain amount of control over the very decisions that affect how much value they can build. 
When Evan sold Blogger to Google, he agreed to work there for two years, but he soon became 
disillusioned. Google hired most of Blogger's team as contractors rather than employees, and it didn't 
come through with all the resources Evan had expected. When Google went public less than two 
years after the acquisition, Evan cashed out some of his stock and left the company shortiy thereafter. 
Similarly, when Dick Costolo sold his third startup, Spyonit, to 724 Solutions, he found he hadn't 



anticipated the potential downsides of his three-year earn-out clause. As Dick explained, 
"Unfortunately, we did the deal in September 2000 and, right after that, the stock deteriorated. Also, 
after the sale, we weren't in control of our company any more. We were no longer responsible only to 
ourselves. There was now a big conpany in Canada that we had to report to. They had very different 
priorities. It was a drastic change." 

Acquisition by a larger conpany may force the startup team that has spent years pursuing its vision 
unhindered to sacrifice that vision if the new mother company has interests that conflict with it. After 
Google bought FeedBurner, Dick Costolo and his cofounders expected their execution risks to drop, 
only to find them replaced by new risks. "We learned that selling the company doesn't de-risk all of 
the execution risk. If you think, 'I'm going to sell because the money's good and they have all these 
resources and I'll make my baby into this beautifiil thing,' sometimes [a sale] actually increases risk 
because of the parent company's competing interests." 

For the Lynx founders, accepting an earn-out would conflict with their major reason for exiting in 
the flrst place: They were burned out after years of riding the entrepreneurial roller coaster, and they 
wanted out. They therefore accepted the offer from Mobilink — millions of dollars lower than 
Spotlight's offer — ^because Mobilink wanted to complete the sale immediately and replace the 
founders. "For us personally, their approach was attractive," James explained. "We wanted to move 
quickly and we really didn't want to stick around for two more years to run things. We were tired and 
needed a break." 

Should We Go PuMc? 

Even for the few founders who are able to remain in a startup until an IPO, this Promised Land is 
often less attractive than expected, bringing many changes that make life harder. 

The most tangible positive change for the company may be that it is now able to access a new 
source of capital — public shareholders. At L90, founder Frank Addante believed that an inftision of 
capital from the public markets and the chance to have a publicly traded stock currency would enable 
L90 to acquire smaller competitors and thereby expand its services. An IPO would also add 
credibility and engender the trust of larger accounts, enabling L90 to bring in new customers. 
Candace Kendle, founder-CEO of life sciences startup Kendle International, Inc., made a similar 
observation: "The number-one best thing is having the public currency with which to complete 
acquisitions. Otherwise we wouldn't have gone that [IPO] route. It's about having money and getting 
more money." 

For founder-CEOs, however, going public can have adverse effects. Academics have theorized that 
having broadly dispersed shareholders after going public might increase the founder-CEO's control 
of the startup,^^ but newly public CEOs often lament the scrutiny and tougher regulations, not to 
mention the substantial cost of mandatory financial disclosure. Frank Addante reflected, "When 
you're private and you take investor money, you have a finite set of investors to manage. You get to 
decide who you work with. You look at them eye to eye and get to know whether they believe in your 
vision, so they share your expectations. When you're public, you don't know who is buying stock. 
You have people who don't understand your ftill vision and are looking at it for a variety of factors. 
Some are short-term, some are long-term, some are buying into the metrics and the fiindamentals and 
some are simply buying based on vision." Kendle International's stock suffered from many such ups 
and downs. Candace Kendle observed, "If you miss your numbers . . . even on a quarterly basis, there 



is a huge price that you pay." 

For Kendle's other cofounder, Chris Bergen, going public had more personal implications. "There 
is a benefit to the public exposure," he explained. "We enjoy a position in the business 
community . . . which we wouldn't otherwise enjoy and that's good. There's a downside to that: Our 
finances are more public than they once were. To the extent that people know how much money you 
make and how much stock you own, that's not something that we're comfortable with because we're 
not very public people, we're very private people." Within Kendle, going public had implications 
further down the organization chart, too. Chris said that, at first, "Employees were absolutely 
ecstatic. . . . [But later], when [the options] are underwater, they think maybe it's not such a great 
deal. Everyone's happy when things are going well; the trick is to make them happy when things 
aren't going well." 

The IPO is much more of an exit for VCs than it is for founders. Not only do VCs get returns from 
their investment, but one study showed that 62% of VCs leave the board once a portfolio company 
goes public. (As a result, many founder-CEOs of newly public startups are answering to new 
boards of directors.) For the founders, as for the startup itself, an IPO is sometimes less an exit event 
than the next in a series of financing events. The founders' previously illiquid stock is now liquid, yet 
it is only after a multimonth post-IPO "lockup" period that founders are able to sell some of their 
stock and enjoy financial gains from their hard work. Even after that, they are subject to constraints on 
how much they can sell; large sales might send a negative signal about the company's prospects, 
harming its stock price. For some founders, there may also be self-imposed constraints on selling 
their stock. Frank Addante, for example, had trouble getting himself to sell his stock in L90: "There's 
a conflict in your mind between selling the stock and holding onto it. It makes you feel that you're 
selling because you believe that you can't create more value in it. It's an emotional struggle 
internally. . . . Anytime I thought about selling some stock, I was never able to pull the trigger." 

Divergence within the Team 

Exit dilemmas can be hard enough to resolve when the founding team is aligned in its motivations; 
they become even harder to resolve if the founders' motivations diverge. For instance, some founders 
may be more risk-averse than others, preferring to take the safe option (such as a "sure" acquisition 
offer) rather than taking a chance on a risky future (such as building more value before selling). Dick 
Costolo explained how risk tolerances diverged and split the Feed-Burner team into two camps: 
"With the different risk profiles within the founding team, some are ready to do the deal, while others 
say, 'Let's tell them to go to hell. We need 25% more.' You might think you're all going to be on the 
same page, but when the real number comes in, people [will be tempted to take what's on the 
table.] . . . Eric and I are always more willing to take more risk, leave it all out there, keep going. 
Steve and Matt are more like, 'Let's not look back on this with regret. Let's take something off the 
table.' " 

Likewise, if one founder is wealth-motivated and the other control-motivated, the former may want 
to grab an offer while the latter founder resists. Indeed, whether or not to exit was the only major 
disagreement between the cofounders of a financial- services outsourcing company. One wanted to 
take a buyout offer from a large public competitor, but the other resisted. The latter cofounder later 
said, "I did everything I could to stop the sale, but eventually realized that the rock was falling 
anyway and that it would be better to be on top of it than under it. It is hard to say that taking $250 



million cash is a bad decision, but [the startup] was my baby and I loved it.""^ 

Early decisions about equity splits may also set up the team to disagree about the best timing for an 
exit. Dick Costolo explained how members of DKA's founding team had different views on their 
startup's sale: "It was not a home run, so there was a lot of grumbling. [And] because we had 
differing amounts of equity, we had differing levels of desire to exit and that caused a real divergence 
within the team." Key nonfounding employees can also hold up the acquisition if they are not aligned 
with the rest of the team Dick described his dilemma when an employee's compensation package 
threatened to derail FeedBurner's sale to Google: "One early employee had a funky option deal that 
won't work out with the way the deal is done, so he doesn't want it. But Google was saying, 'That 
person's a key employee,' so that person can hold the deal hostage." 



* In Hall et al's (2010) study of venture -backed companies from 1987 to 2008, of the startups that had exited (half of their sample), 
approximately 52% had merged or been acquired, 18% had gone public, and 30% had been liquidated or had gone bankrupt. 

* As Shakespeare stated (in As You Like It, 2:7), "All the world's a stage/And all the men and women merely players;/They have 
their Exits and their Entrances." 

* These names are pseudonyms. 



ACKNOWLEDGMENTS 



Writing a book is a team sport . Even if the player is the one with his picture on the baseball 
card, he would not have achieved the statistics on the back without a squad of coaches, scouts, and 
support staff. Similarly, without my coaches here at Harvard Business School and at other schools, 
the scouts who decided to place early bets on me (including Seth Ditchik, who was willing to bet on a 
rookie author), and the people who provided support for every phase of this book-writing process, 
this book would never have made it up to bat. 

It was Nitin Nohria who sparked my interest in an academic career and, for more than a decade, 
has provided mentorship and support on all levels. Paul Confers, Josh Lemer, Bill Sahlman, George 
Baker, Jerry Green, and Peter Marsden were invaluable early research mentors who encouraged me 
to focus on rigorous inquiry into the black box of founding teams. Lynda Applegate and Teresa 
Amabile not only have mentored me ever since I was a rookie in the recently created Entrepreneurial 
Management unit, but also became deeply engaged with the early drafts of this book, helping shape its 
final state in a myriad of ways. During the same stretch, Tom Eisenmann provided terrific career 
guidance, conpanionship through many lunches, and targeted feedback on how to improve this book. 
Toby Stuart, Geoff Jones, Nancy Koehn, and Kash Rangan were very early and influential supporters 
of the idea of writing a book to tie together a decade of research. Ben Esty, David Garvin, Andre 
Perold, Bharat Aland, Ranjay Gulati, Mihir Desai, and Jan Rivkin provided sagacious input during 
my early days on the faculty of Harvard Business School, while I was developing the course on 
which The Founder s Dilemmas is based and at other key junctures of my career. Further afield, but 
still foundational, are my debts to Howard Aldrich, Donald Hambrick, Kathy Eisenhardt, and 
Howard Stevenson, four of the giants on whose shoulders my research stands. 

The seeds for the book's data were planted early and have had several key nurturers. Brian Hall 
facilitated my initial dialogue with Bill Holodnak and Aaron Lapat about a potential collaboration. 
Bill and Aaron in turn trusted a young doctoral student to write the CompStudy survey instrument and 
analyze its results. Mike DiPierro and Evan Brown have kept everything moving smoothly year after 
year, even when they found themselves having to keep more balls in the air than we expected. Furqan 
Nazeeri, an invaluable collaborator on multiple projects, was a terrific sounding board for n^ early 
research. 

A wide variety of collaborators helped me craft the cases and papers described in this book. Thank 
you to Deepak Malhotra, Henry McCance, Eric Olson, LP Maurice, Antony Uy, Rachel Gordon, 
Rachel Galper, and Rosy Fynn for coauthoring the case studies that form the backbone of my MBA 
course and of the book; to Jeff Bussgang, not only for coauthoring a case but also for the plethora of 
subsequent projects and collaborations that case has sparked, and also for making a wonderfiilly 
insightfiil pass through an early draft of this book; and to all of my case protagonists for their 
openness and honesty in helping me learn about the toughest dilemmas they had faced and for their 
help in crafting case materials that helped share their knowledge with the next generation of high- 
impact founders. Thank you to all of my Entrepreneurial Manager and Founders' Dilemmas students 
(and Kaufiman Scholars) for continually showing me new sides of our material (two of which 
resulted in research projects that would not have existed otherwise) and for giving me firsthand 
appreciation for R' Chanina's ancient Talmudic saying, "I have learned much wisdom fi"om my 



teachers, more from my colleagues, and the most from my students." 

On the research side. Matt Marx, Thomas Hellman, and Warren Boeker have been delightful 
collaborators, often at all hours of the night, on three of the papers profiled in this book. Matt Marx 
also believed in the Founders' Dilemmas course enough to stake his early MIT teaching career on 
becoming my "first franchisee." Wendy Torrance likewise believed in the inportance of the material, 
making it a central part of her Kauffimn Global Scholars program Thank you to Matt and to Ginger 
Graham for helping make teaching a delight as our weekly three-professor teaching group 
brainstormed about the material we would be covering and the best ways to educate our future 
founders about the dilemmas awaiting them 

Li addition to being wonderful colleagues throughout the years, Ramana Nanda, Mukti Khaire, Bill 
Kerr, Monica Higgins, Rakesh Khurana, Boris Groysberg, and David Scharfstein provided key input 
at various points in the research and book-writing process. Likewise, Bill Schnoor, Janet Kraus, Tim 
Butler, Evan Richardson, Tim Connors, Amar Bhide, Tony Mayo, and Ari Ginsberg provided well- 
timed, high-inpact input on specific parts of the book. 

Thank you to Lisa Brem and Kyle Anderson for the depth and breadth of their contributions to this 
book, sharpening the final product in so many ways while making the writing process lots of fun, and 
to John Elder for his input into every piece of every chapter. Day in and day out, Theresa Gaignard 
and Matthew O'Connell provided supremely conpetent and enjoyable administrative support. 

Finally, acharonim chavivim, thank you to my eishet chayil, Ghana, for being an inconparable 
wife and mother; to Talya, Tamar, Yair, Liat, Naava, Avital, and Yishai for amazing and amusing me 
every day with their accomplishments, enthusiasm, and delightful personalities; to my wonderful 
parents and in-laws for laying the foundation on which three generations have been built; and to 
Hakadosh Baruch Hu for making it all possible. 



APPENDIX A 

QUANTITATIVE DATA 



This book's quantitative data come from a unique dataset of thousands of startups, founders, 
and executives collected over the last decade, from 2000 to 2009. This appendix describes the survey 
process that was used to collect the data, the demographics of the respondents, and the contents of the 
survey instrument. 

SURVEY PROCESS 

When I started doing systematic research on founders and high-potential startups in 1999, one of the 
major challenges I faced was the lack of publicly available data. In contrast to public companies, 
which have to publicly file detailed disclosures about compensation, company operations, executives, 
and other important aspects of the company, privately held companies can keep all of those details to 
themselves. There is no private-company version of the SEC's Edgar database. The few databases 
that have tried to compile such data are extremely limited in their scope and plagued by missing-data 
problems. In particular, startups are extremely secretive about executive compensation,^ 
capitalization tables, financing valuations, founder-CEO succession events, and founders' equity 
splits. There are no good sources even for some less-sensitive data points, such as the founders' prior 
relationships and pre-founding backgrounds. Most research on high-potential startups has therefore 
either used field-oriented approaches or focused on companies that had recentiy gone public (at 
which point the founders rarely still number among the top executives). For instance, a seminal 1994 
study of "young company" compensation that used data on newly public companies was not able to 
examine founder condensation.^ Researchers who have tried to examine founders in large companies 
discovered that founders made up less than 10% of the executives in their datasets, precluding 
definitive founder-related conclusions.^ 

To solve this problem, I started collecting my own data using a survey of private high-potential 
startups. In 1999, I partnered with three national professional-services firms — Ernst & Young (an 
accounting firm), J. Robert Scott (an executive- search firm), and Hale and Dorr (a law firm) — to 
conduct our first survey of technology startups. I designed the survey instrument, pilot- tested it with 
10 companies, and revised it based on their feedback and my analyses of their responses; the three 
partner firms developed a website to administer the survey, which we named "CompStudy" (a name 
we extended to the survey and to the reporting site, www.compstudy.com); and I analyzed the data 
and compiled detailed breakdowns of the results. For the first decade of the survey, we focused on 
the United States, given its important role in entrepreneurship and our intimate knowledge of the 
country, but we are now expanding the survey to four other countries in which entrepreneurship is 
either strong or growing — the United Kingdom, Israel, China, and India. 

Given the breadth and sensitivity of the survey questions, we decided to target the CEOs and CFOs 
of the startups, with the expectation that a single senior executive would conplete the entire survey 
for each company.* To provide incentives to those executives, we interviewed CEOs and board 
members of startups and learned that they lacked solid data on executive compensation. (They were 



facing the same lack of data that we faced as researchers!) As a result, they did not know how much 
salary, bonus, and equity to offer potential hires and were losing many battles for that talent; they 
faced problems retaining their own staff because they could not accurately gauge what the market was 
paying for executives; they feared that they were overpaying some executives even though the typical 
startup is cash-poor and can't afford to overpay; or they found that they had given away too much 
equity to their hires and were left with little themselves. 

As a "carrof to drive participation, our research team promised participants that we would 
conpile the survey's condensation data into a detailed "Compensation Reporf that would slice the 
data for all C-level and VP-level positions by founder/nonfounder status, company maturity, industry 
segment, geographic region, and other dimensions. We would provide the report only to participants 
who completed the survey — at no charge to them To help ensure valid and complete submissions, we 
built data- validation capabilities into the online survey. 

Over the past decade, the annual Compensation Report (commonly referred to as the Ernst & 
Young Compensation Report or the Harvard Business School Compensation Report) has become 
highly visible within the industry, serving as a standard benchmark for startups making hiring and 
retention decisions and deciding how much to budget for compensation and equity stakes. The data 
have also served as the backbone of a series of academic papers.^ 

The initial invitation list to participate in the survey was derived fi^om multiple listings of 
technology startups across the country, including the membership lists of regional and statewide 
technology councils, the VentureOne database of companies that have raised venture capital, the 
client lists of the professional-services firms with which I was partnering, and recommendations by 
private-company investors. As will be described below, respondent locations closely matched the 
distribution of high-potential startups across the country rather than being concentrated in just one or 
two regions. Regular tests of geographic location and of startup maturity and size have shown no 
statistically significant differences between respondents and nonrespondents; details are provided 
below and in the references.^ 

Analyses of the first three years of responses indicated that a 20% response rate was achieved 
over those years,^ a relatively high response rate considering the sensitivity of the questions and the 
level of the executives targeted.^ Those calculations were possible due to our having a clearly 
defined list of startups to which we were mailing survey invitations. Since 2004, though, the 
invitation approach has been augmented in two ways: We developed relationships with professional 
investors who forwarded survey invitations to all of the companies with which they were involved, 
and we contacted respected bloggers who posted the invitations for their readers and subscribers. 
These changes have made it harder to calculate precise response rates, since we cannot know how 
many legitimate candidates for the survey received the forwarded emails or saw the blog posts. 
(These changes have also increased our need to validate each response.) However, the growth in the 
survey's visibility, reputation, and usefiilness would suggest that response rates were probably 
maintained fi"om the rate achieved when the survey was in its initial years. 

Any database of currentiy operating companies is susceptible to survivorship bias. For instance, 
the databases of public companies that are commonly used in academic research include only 
conpanies that survived long enough — and succeeded well enough — to go public, but there are 
systematic differences between those companies and the broader population of all companies. 
Although the CompStudy dataset is much less susceptible to such biases because it catches startups 
soon after founding, some CompStudy analyses may be susceptible to survivorship bias. Whenever 



possible, I have used methods and analyses to assess the level of potential bias or to correct for it, or 
else I report differences by stage of startup evolution. For example, as described in Chapter 11,1 
performed two separate analyses — one using the younger half of startups and another using the older 
half — to analyze the trade-off between the value of founders' equity stakes and the degree of control 
the founders kept, and then compared and reported the results. 

Over the years, the percentage of repeat respondents has increased from almost none to about 
20%.* In analyses that would be sensitive to autocorrelation problems, or where only one 
observation was desired per startup (e.g., when doing analyses of founding teams, wherein repeat 
respondents would have multiple identical entries for each founding team), ' multiple methods were 
used to avoid such problems, including dropping repeat respondents, recalculating robustness models 
to see if the results changed, and including fixed effects in the models. 

DEMOGRAPHICS OF RESPONDENTS 

The 2000 and 2001 surveys targeted technology startups. Those two years, we received complete 
responses from 211 and 178 startups, respectively. In 2002, to enable cross-industry conparisons and 
enable us to cover both of the major industries for high-potential startups, we broadened the survey to 
include life sciences startups and received conplete responses from 168 of them. In subsequent years, 
the number of responses grew steadily, though more quickly in the technology industry than in life 
sciences. From 211 startups in its first year, the technology survey grew to include 489 startups in 
2009; the life sciences survey grew from 168 startups in 2002 to 214 in 2009. 

The frill decade of data include a little more than 3,600 startups (32% from life sciences, 68%) from 
technology), 9,900 founders, and more than 19,000 executives. The distributions of startups regarding 
location, financing stage, number of employees, and age of startup was as follows: 

• Geography — The geographic distribution of respondents closely matches accepted profiles of 
the startup population, with the two biggest entrepreneurial hubs making up the two largest 
segments of CompStudy respondents. For the decade, 32%) of CompStudy startups were in 
California, 22% in New England, 19%o in the Mid-Atlantic, 15% in the western United States 
(excluding California), and ll%o in the southern United States. These percentages were almost 
identical for the technology and life sciences indusfries, with no region differing by more than 
2 percentage points. 

• Financing — The respondents to the first technology survey were almost all VC-backed, 
matching the disfribution of startups in VentureOne and other standard databases that include 
only startups that have received institutional financing. Over the years, the percentage of 
startups that had not raised any outside financing grew from 2%) to 11%) (12%) in technology, 
9%) in life sciences). Across the decade, 7%o had not raised any financing, 18%o had raised one 
round, 28%) two rounds, 23%o three rounds, 12%o four rounds, and 12%) five or more rounds. 
This suggests that the sample is biased toward founders who are willing to take outside 
financing; wealth-motivated founders seem to be overrepresented. To the extent that confrol- 
motivated founders are underrepresented, the quantitative results should understate the 
strength of any wealth- versus-confrol differences; those differences should be even more stark 
with a dataset that has more control-oriented founders. 

• Number of employees — On average across the decade, the startups had 28 fiiU-time- 



equivalent employees (PTEs). In the technology industry, 34% of the startups had 1 to 20 
PTEs, 28% had 21 to 40 PTEs, 20% had 41 to 75 PTEs, and 19% had more than 75 PTEs. 
Life sciences startups were smaller, with a little more than 50% of the startups having 1 to 20 
PTEs and lower percentages than technology startups in each of the other buckets. 

• Age of startup — The average age of the startups in the full dataset was 6.8 years, with a 
standard deviation of 4.4 years. The 25th percentile was 3.7 years, the median was 6.1 years, 
and the 75th percentile was 9.0 years. Of the startups, 84%) were founded between 1996 and 
2006. 

SURVEY CONTENTS 

The core survey questions cover conpany founding, dates on which key product-development 
milestones were achieved, financing history, backgrounds of the members of the executive and 
founding teams, executive compensation, and the conposition of the board of directors. Note that 
these questions take a "revealed preferences" approach to understanding founders; rather than asking 
founders about their preferences and inclinations, we focus on their actual decisions and actions as a 
more revealing window into their preferences and inclinations. The same core questions have been 
used (with only minor refinements) since the first year of the survey, but some sections have been 
expanded since then. In particular, in 2006, the founding-team section was expanded to include 
detailed questions about founders' equity splits and their prior relationships. Those questions play a 
central role in Part n of this book. In 2008, the section on founder-CEO succession was expanded to 
include more questions about the trigger of succession, the replaced founder-CEO's post-succession 
role, and other questions that are used in Chapter 10. 

The specific questions from each major section of the survey are as follows: 

• Por each company: 

c Date company founded 
o Industry segment, location 

c Date on which development of the initial product or service was conpleted (or is expected 

to be completed) 
c Company revenue, whether it is profitable, headcount 
c Average monthly burn rate, cash on hand 

c Capitalization table: percentage of equity owned (frilly diluted) and percentage of equity 
owned (current) for: 

■ Current enployees, former employees 

■ Angel investors, VCs, corporations/strategic investors 

■ Option pool for friture hires 

■ Other equity 

c Number of CEOs 

c Current CEO's frmctional background 



Por each founder: 



o Prior experience 



■ Whether founder had previously founded another company 

■ Years of work experience before founding this company 

■ Whether founder had prior experience managing people 

c Founder(s) whose idea it was to begin this venture 

c Initial position in the company 

c Whether founder was working full-time for the startup from the day it was founded 

c Amount of founding capital contributed by this founder 

c Percentage of company's equity received at time of initial equity split 

c Whether founder is currently employed with the company 

■ If not currently employed with the company, number of months this founder worked as a 
full-time employee in this company 

• For each founding team: 

c At the time of founding, how many months had any of the founders already worked on this 

project with the intention of founding a company? 
c Prior relationship within the founding team: Before founding this company . . . 

■ How many of the founders had previously worked together? 

■ How many of the founders had founded another company together? 

■ How many of the founders were friends but not coworkers? 

■ How many of the founders were related to each other? 

c Equity split process 

■ Date the founders initially split the equity 

■ How much time the founders spent negotiating the initial equity split 

■ When the founders split the equity, whether they did so informally (i.e., not in writing) 
or formally (in writing) 

■ If a founder had left the company soon after the equity was split, would the founder have 
had to give up a significant amount of his or her equity? 

■ When the first institutional round of financing closed, what percentage of the founders' 
equity was fully vested? 

• For each round of financing: 

c Date round closed 
c Pre-money valuation 
c Gross equity proceeds 

o Participants in the round: founders, angels, VCs, corporate/strategic investors, debt 

providers, public equity markets 
c Number of VCs participating in the round 

c Liquidation preferences: none, fx, l.lx-2x, 2.1x-3x^ more than3x 



• Sources of equity investment leads: How company found initial angel investor, subsequent 
angel investor(s), initial VC, subsequent VC(s), initial corporate investor, subsequent 
corporate investor(s) 

c Friend, family, or past coworker of the CEO; friend, family, or past coworker of other 
team member; investor in past company of a founder; through submission of business plan 
or cold call; referred by existing angel investor; referred by existing VC or member of 
board of directors; referred by investment bank or other outside professional (e.g., lawyer, 
accountant, commercial bank); other source 

• For each executive (C-level or VP-level): 

c Prior experience 

■ Number of years of fiill-time career work experience before joining company 

■ Academic degree(s) 

■ Before joining this startup, was this person ever before the senior-most executive within 
his or her function? 

c Date joined the startup 

o Source of hire: Prior connection to CEO (e.g., past coworker, friend, family member), title 
of another member of the executive team with whom this person had a prior connection, 
referred by existing investor or member of board of directors, other source 

c Whether founder or nonfounder 

c Gender 

c Condensation 

■ For previous and current years: annual base salary 

■ For previous and current years: cash bonus (planned and received) 

■ Percentage of current-year bonus that is guaranteed 

o Equity holdings: 

■ Percentage of company's total outstanding equity currently held (Mly diluted) 

■ Percentage of company's total outstanding equity granted at time of hire (frilly diluted) 

■ Equity vehicles used at time of hire: incentive stock options (ISO), nonqualified stock 
options, resfricted stock, common stock 

■ Vesting: Number of years over which frill vesting occurs, whether vesting is 
performance based, whether executive would have accelerated vesting upon change of 
confrol 

■ Frequency of equity grants afrer time of hire 

c Whether executive was eligible to receive a severance payment if fired and, if so, number 
of months of severance 



• For each board member: 



c Date joined board of directors 
o Background 

■ If current employee, title in company 

■ If former employee, last title in conpany 

■ Whether angel investor, VC, executive in same industry, executive from outside the 
industry, academic, other 

c Compensation 

■ Equity stake: percentage of frilly diluted shares granted to join board, currently held 
percentage of total outstanding shares (frilly diluted, including both stock and options) 

■ Annual equity retainer: percentage of total outstanding shares (frilly diluted) 

■ Annual cash retainer 

■ Committees: serves on audit committee, executive committee, compensation committee, 
or other committee 

• For each company that had replaced the founder-CEO: 

o First nonfounding CEO's background 

■ Years of prior experience 

■ Functional background: 

■ Whether had prior work experience in the industry, as a C-level executive, in a private 
company, in a public company 

o If first nonfounding CEO has been replaced, for how many months he or she remained CEO 
of the company 

c Date when decision was made to replace the founder-CEO 

c Main trigger of the change in CEO: board, founder-CEO, board and founder-CEO equally, 

non-CEO enployee, other 
c After being replaced, did the former founder-CEO remain at the conpany in an executive 

position? 

■ If yes, which position: CTO, other C-level position, lower-level executive role 

c After being replaced, did the former founder-CEO remain on the board of directors? 

■ If yes: became chairman, remained chairman, moved from chairman to non-chairman 
director, remained non-chairman director 

c Annualized percentage sales growth 

■ In the 6 months before the founder-CEO was replaced 

■ In the 12 months after the founder-CEO was replaced 



* Survey respondents consistently report spending an average of one hour completing the survey. 

* A low rate of repeat survey participation is not surprising, given the high rate of failure among startups, demands on the time of their 
CEOs and CFOs, the exclusion of startups that had gone public or been acquired, and other factors. 

^ The increase in repeat respondents has also enabled us to test the reliability and accuracy of survey submissions by comparing the 
answers from repeat respondents. 



APPENDIX B 

SUMMARY OF STARTUPS AND PEOPLE 



The two tables below present alphabetical listings of the startups described in The Founder s 
Dilemmas and of the people (mentioned in the book) who founded or worked in them. The startups 
table also includes references to the full case studies that have been published about those startups. 

Table B.l 

Table of Startups 



Surtup Nitmf 


Yfjr 

founded 


ProUigomst 
ihiilul Titif 
in Sutrlup) 


Other titiittding 
Teitm Memtterf 
llnitui Titlei) 


Prior 

Relationship 
with 

Cofoimderfs)/ 
Partnerts} 


Mentioned 
in Chapters 
Ihrief 

mentions in 
pjrenthes^) 


Origiiul Idej for Startup 


is Studios' 


2006 


Curt Schilling 
(Founder- 
President- 

( K iirm in t 


N/A (solo 
founder) 


N/A 


2,(11) 


Develop a best-of-brecd 
online role-playing 
g.inie with a rich Mory 


Apple 
Computer* 


1976 


Steve W'ozniak 
(Cofounder) 


Ste*t Jobs 
(Cofounder) 

Ron Wayne 
(Cofounder) 


Close friend 
Stranger 


4, (5). (6), 
(7) 


Develop and sell a 
personal computer based 
on VTomiak's design. 


Bloggcr' 


J 999 


E»jin VTiiiiaim 
(Founder-CEO) 


Meg Hourihan 
(We President) 


Ex -girlfriend 


1.4.5.6.7, 
8, 9. 1 1 


Online self-publishmg 
(hlogging) tool. 


Cirdn 


1997 


J^nct Krjiui 
(CVtfounder- 
CEO) 


Kalhy 
Sherbrooke 
(Cofounder, 
President/coo » 


Fellow MBA 
snident 


(I)a4),(6) 


Provide "everyday 
services" to individiiaU. 
such as dog walking, 
errands, etc. 


ConneXui' 


1996 


Humphrey 
Chen 

(Cofounder) 


CeorKe Searle 

(Cofounder) 


Fellow .MBA 
student 


2 


Develop a network of 
coiiiputers that will 
automatically catalogue 
iiuiMC playing un 
the radio and allow 
consiuners to access song 
information, listen to the 
song, and purchase it via 
a toiichtone phone or 
Internet website. 


Digital 
Knuw lodge 
Aisrts (DKA)' 


1995 


Dick Costolo 

(a>-VPof 

Technology! 


Eric Lunt 
(a>-VPof 
Technolog)') 
Two other 
founders (CEO, 
COO) 


Coworker 
(suKirdiiute) 

Strangers 


(S) 


Collaboration and 
publishiitg software 
company. 


FeedBiiroei* 


2003 


Dick Costolo 
(Founder-CEO) 


Eric lAint 
(CrO) 


Cofounders 


1,(3). 5. 7. 
8.9.11 


Intermediary between 
online publishers. 



advertisers, and 
distributors. 

Matt Shobe 
(ProdiKt 
Design) 
Steve 

Olechowski 
(COO) 



Kcndle 

lntcrnarioii.ll 


1981 


C.and.Ke 
Kendle 

(Founder-CEO) 


Christoph<T 
Bergen 

(President and 
COO) 


Colleague. 

fiancee 


Appendix 
11-2 


To build a clinical 
research organization 
(CRO) to perform 
clinical driig trials 
for pharmaceutical 
companies. 


L90' 


1997 


Fr.ink AJJantr 
(Foimder-CTOl 


John Boh.in 
(Foiuider C KO) 


Busiitcss 
contact 


(6). (8) 


Merge ad «-rvcr with ad 
broker «> take advantage 
of syi>ergies. 


Lynx 

Solutions*'* 


1998 


•James Milmo 
(Foiuider- 
President- 
Chairman) 


•Doug Curtis 
(CEO) 

•Javier Pascal 
(CTO) 


Acquaintance 
Friend 


2. 5. 6. 8. 9, 
10. 11 


Advertising-supported 
screensavers for handheld 
wireless devices. 


MaicrRy'* 


2000 


Barr> NalU 
(Founder-CEO) 


N/A (solo 
founder) 


N/A 


1.2.3.8. 
(9). 11 


Telecom service provider 
for enterprise and B2B 
customers. 


Mcgaser»'er"* 


-1996 


Les Trachinun 

(profesuon.ll 

CEO) 


Mother (Ch.iir 
of hoard) 

Father 
(Inventor/ 
technologist) 
Soil (COO. 
former founder- 
CEO) 


Family 
nieinhers 


10 


Enable multiple 
cmiputing platforms to 
form a supercomputer. 


Nike" 


1964 


Phil Knight 
(Founder- 
President) 


Bill Bowemian 
(We President) 


Coach 


(6) 


Create and market an 
innovative running shoe. 


Ockham 
Technologies" 


1999 


Jim Trundiflou 
(Founder-CEO) 


Mike 

Mrisenheimer 
(\'P Product 
Management) 
Ken Burows 

(COO) 


Coworker 
(subordinate) 

Coworker 


2. (3). 4, 5. 
6,7,9,(11) 


Sales-force-opnmization 
software. 


Odeo'* 


2004 


Evan Williams 
(Cofounder- 
investor- 
adviior) 


Noah GkkM 
(CEO) 


Acquaintance 


1.(2). 4. 5. 
6.8.9,11 


PoJcastuig technology. 


l*.iiia.)ra" 
(originally 
Savage Beast) 


2000 


Tim Westcrgxen 
(Founder-Chief 
Musk Officer) 


Jon Kraft 
(CEO) 

Will Glaser 
(CTO) 


AcquainiaiKe 

Stranger 


1.2.(3). 4. 
5, (6), (7). 8. 
9,11 


A mUMC deciMon engine 
that plays/suggests songs 
based on attributes of 
songs/artists the user 
chooses. 


Proteits 
Biomedical " 


2001 


Andrew 

Thompson 

(Founder-CEO) 


George Savage 
(CMC) 

Mark Zdehlick 

(CTO) 


Fellow MBA 
student 

Acquaintance 


2 


To develop technology 
that could insert 
miniature computers and 
sensors into drugs and 
medical devices. 


Rea.\i««is'' 


1997 


Frank Addante 
(Ciofounder) 


"Cary" 
(Cofounder) 


Friend/ 
colleague 


6 


Advertising software 
that tracks and xhedulrs 
a company's online 
adi'ertising. 


RuhhisI) B..^^'" 
(renamed 
l-800-GOT- 
JUNKl 




Brian 

Scudamore 
(Founder) 


"John" 
(Cofounder) 


Friend 


(2).i3).9. 
II 


Trash-renioval service. 


ScRwjy'* 


1999 


Dean Kamen 
(Founder) 


N/A (solo 
founder) 


N/A 


10. 11 


Create a low-cost, 
low -emission personal 
transportation device. 


Sittercity^ 


2001 


Genevieve 
Thiers 

(Founder-CEO) 


N/A (v>lo 
founder) 


N/A 


1.2.4. 8. 

10, 1 1 


An online inalchiiuking 
service for parents and 
babysitters. 



Snurtix" 


1999 


Vwek Khuller 
(FoiindcT-CEO) 


Kinll Dmitriev 
Saurahh Mittai 

Unnamed 

technical 

ccifounder 


Fellow MBA 
students 

Coworker 
(subordinate) 


1,2. (3), 4. 
5,6.7 


Create an electronic 
ticketing technology for 
sports and entertainment 
venites. 


Spyooir* 


-1998 


Dick Costolo 
(Founder-CEO) 


Eric Lunt 

(cro) 

Matt Shobe 
(ProdiK-t 
[)esignl 
Steve 

Olechowski 
( Business 
Det'elopinenc/ 
Programmer) 


Cofounder 

Coworkers 
(subordinates) 


(3), 8 


Alert systems for Internet 

consumers. 


SirongXtail" 


2002 


Frank Addanre 
(Founder-CEO) 


Tim McQuillen 
(Cofounder) 


Coworker 

(subordinate)/ 

friend 


2.7.9,11 


Create an email 
messaging infrastructure 
software package and sdl 
to enterprise clients. 


Trjfwconrive'* 


1981 


Lcs Trachtnun 
(VT 

Operations, 

professional 
CEO) 


Mike Brody 
(Founder-CEO) 

Mike's brocher 
(Founder-CFO) 


Mike's family 
members 


10,(111 


Corporate stock-option 
software. 


UpDown" 


2007 


Michael Reich 
(Founder-CEO) 


Georg 
LudviksMHi 
(Co-CEO) 
Phuc Truong 
(CTO) 


Fellow MBA 

students; 

stranger 


3, (4). 6, 7 


Create an online social- 
networking site for 
investors. 


Wily 

Technology^ 


1998 


Lew C irne 
(Foiuider-CEO) 


N/A (5t>lo 
founder) 


N/A 


(3), 8. (9), 
10. II 


Self-diagnmtic software 
for computer operating 
sys tents. 




2000 


Robin Chase 
(Founder-CEO) 


Antje 

Danielson (NT- 
Environmental 
attairs and 
strategy) 


Friend 


(2), 6, (8) 


Create a car-sharing 
business based on a 
German company and 
nurket it in the United 
States. 


ZondiRo" 


2000 


Frank Addante 

(Founder-CEO) 


Two unnamed 
cofounders 


Business 

contacts 


5, 6. 8. 1 1 


Create technology' for 
"on-the-go" advertising 
dirivt to wireless de%-ice*. 



^ Disguised name. 

^ Noam Wasserman, Jeffrey J. Bussgang, and Rachel Gordon, "Curt Schilling's Next Pitch," HBS No. 810-053 (Boston: Harvard 
Business School Publishing, 2010). 

^ Noam Wasserman, "Apple's Core," HBS No. 809-063 (Boston: Harvard Business School Publishing, 2009). 

^ Noam Wasserman and LP Maurice, "Evan Williams: From Blogger to Odeo (A) and (B)," HBS Nos. 809-088 and 809-093 (Boston: 
Harvard Business School Publishing, 2008). 

^ Monica Higgins and Noam Wasserman, "Humphrey and Cecilia," HBS No. 810-702 (Boston: Harvard Business School Publishing, 
2010); Monica Higgins, Adam Richman, and John Galvin, "Video Case: Humphrey Chen (Preview Note)," HBS No. 498-036 (Boston: 
Harvard Business School Publishing, 1997). 

^ Noam Wasserman and Eric Olson, "Lather, Rinse, Repeat: FeedBurner's Serial Founding Team," HBS No. 809-089 (Boston: 
Harvard Business School Publishing, 2009). 

^ Noam Wasserman and Eric Olson, "Lather, Rinse, Repeat: FeedBurner's Serial Founding Team," HBS No. 809-089 (Boston: 
Harvard Business School Publishing, 2009). 

^ Dwight B. Crane, Paul W. Marshall, and Indra A. Reinbergs, "Kendle International Inc," HBS No. 200-033 (Boston: Harvard 
Business School Publishing, 2000). 



Noam Wasserman and Antony Uy, "Frank Addante, Serial Entrepreneur," HBS No. 809-046 (Boston: Harvard Business School 
Publishing, 2008). 

^ Noam Wasserman, "The Tale of the Lynx (A), (B), and (C)," HBS Nos. 807-151, 807-152, 807-153, (A and B): 807-112 (Boston: 
Harvard Business School Publishing, 2009). 

Noam Wasserman and Rachel Ga^er, "Big to Small: The Two Lives of Barry Nails," HBS No. 808-167 (Boston: Harvard 
Business School Publishing, 2008). 

Noam Wasserman and Rosy Fynn, "Les Is More, Times Four," HBS No. 807-173 (Boston: Harvard Business School Publishing, 
2008). 
1 2 

Noam Wasserman and Kyle Anderson, "Knight the King: The Founding of Nike," HBS No. 810-077 (Boston: Harvard Business 
School Publishing, 2010). 
1 3 

Noam Wasserman, "Ockham Technologies: Living on the Razor's Edge," HBS No. 804-129 (Boston: Harvard Business School 
Publishing, 2004). 

Noam Wasserman and LP Maurice, "Evan Williams: From Blogger to Odeo (A) and (B)," HBS Nos. 809-088 and 809-093 
(Boston: Harvard Business School Publishing, 2008). 

^5 Noam Wasserman and LP Maurice, "Savage Beast (A) & (B)," HBS Nos. 809-069, 809-096 and (Al) 801-051 (Boston: Harvard 
Business School Publishing, 2008). 
1 fi 

Richard Hamermesh, Lauren Barley, and Ginger Graham. "Proteus Biomedical: Making Pigs Ffy," HBS No. 809-051 (Boston: 
Harvard Business School Publishing, 2008). 
1 7 

Noam Wasserman and Antony Uy, "Frank Addante, Serial Entrepreneur," HBS No. 809-046 (Boston: Harvard Business School 
Publishing, 2008). 

Noam Wasserman and Rachel Galper, "Rubbish Boys," HBS No. 808-101 (Boston: Harvard Business School Publishing, 2008). 

Richard Hamermesh and David Kiron, "Managing Segway's Early Development," HBS No. 804-065 (Boston: Harvard Business 
School Publishing, 2003). 

Noam Wasserman and Rachel Gordon, "Playing with Fire at Sittercity (A) & (B)," HBS Nos. 809-009 and 809-010 (Boston: 
Harvard Business School Publishing, 2009). 

Noam Wasserman, "Smartix: Swinging for the Fences," HBS No. 808-1 16 (Boston: Harvard Business School Publishing, 2009). 

Noam Wasserman and Eric Olson, "Lather, Rinse, Repeat: FeedBumer's Serial Founding Team," HBS No. 809-089 (Boston: 
Harvard Business School Publishing, 2009). 
73 

Noam Wasserman and Antony Uy, 'Trank Addante, Serial Entrepreneur," HBS No. 809-046 (Boston: Harvard Business School 
Publishing, 2008). 
24 

Noam Wasserman and Rosy Fynn, "Les Is More, Times Four," HBS Case No. 807-173 (Boston: Harvard Business School 
Publishing, 2008). 
25 

Noam Wasserman and Deepak Malhotra, "Negotiating Equity Splits at UpDown," HBS No. 809-020 (Boston: Harvard Business 
School Publishing, 2008). 
7^^ 

Noam Wasserman and Henry McCance, 'Tounder-CEO Succession at Wify Technology," HBS No. 805-150 (Boston: Harvard 
Business School Publishing, 2007). 

Myra Hart, Michael J. Roberts, and Julia D. Stevens, "Zqjcar: Refining the Business Model," HBS No. 803-096 (Boston: Harvard 
Business School Publishing, 2003). 

78 

Noam Wasserman and Antony Uy, "Frank Addante, Serial Entrepreneur," HBS No. 809-046 (Boston: Harvard Business School 
Publishing, 2008). 



Table B.2 

Table of People 



Protagonist 


Stjrtiip 
\jnif 


TillcitnJ R'Je in Startup 


Highest 
Degree 
ReceireJ 


Wort Experience 


Other Relei\int 
Experience 


Addantr, 
Frank 


RcaXions 
L90 

Zondigo 
StrongMail 


Cotoiindcr 

Product and Inisine^s 

dct'clopmcnt, hired employees 

Negotiated merger to form L90 

Foiinder-CTO 

TechnK'al devch^mrnl 

Founder-CEO 

Business and product 

development 

Secured funding, hired team 
Founder-CEO 
BiiMiK>N and prodiKT 
development 

Secured tunding, hired executive 
team 


High school 


Scl(-taiight 
programmer; Tl line 
mstaller 


Four years of 
college before 
leaving 


Chnitophcr 


Kcndle 
International 


Founder, PreMdent and COO 


MBA 


Associate \'P. 
Children's Hospital of 
Philadelphia 




Bohji), 
John' 


LW 


Founder-tEO 

Secured funding, directed public 
offering 


Economics 
degree 


OpeiictI NXest t <ia»t 
office of early website 
company; T\' ad sales 




Bow«mian. 
Bill 


Nike 


Vice President and director 
Product development 


Four-year 

college 

degree 


Ciollege and Olympic 
track-and-lield coach 


Developed 
running shoes 


Brody, 
Mikc- 


TranKcntivc 


Founder-CEO 




CEO of digital photo 
lab 

Cofounder of 
Stockholder Systems 




Burows. Ken 


(Xkham 
Technologies 


COO 

Business de%°elopment 
Dropped out M founder 


MBA in 
finance and 
strategic 
management 


Systems integration 
consultant 
Director-Financial 
Infornuuion Systems 
Business development 




•C»ry- 


ReaXiiifW 


Cofouitder 
Investor 




IT consultant 




Chasc. 
Robin 


Zipcar 


Founder-CEO 

Refined concept, market research, 
developed business plan 5c 
budgets, designed website 
Hired contract engineers for 
onUne reseo'ation sj^tem 


MBA in 
applied 

economics 
and finance 


Healthcare consulting 
Private school director 
of finance and 
operations 
Managing editor of 
public health |oumal 





Chen, 
Humphrey 


ConneXus 


Cofounder 
Business and product 
dociopment. secured hmding 


MBA 


Five years IT experience 
in several companies. 
iiKluding online music 
store startup 




CiriK, [jew 


Wily 

Technology 


Founder-CEO 
Product and business 
de>-elopment, hired employees, 
early sales tu cu>ttiniers, secured 
funding 


B.A. 

computer 
scieiice 


Senior software 
engineer at Apple 
Computer 
Java project lead 
at Hummingbird 
Communications 




Coitolo, 
Dick 


DKA 
Spyonit 

FccdBumer 


Co-VP Teihuology 
Product iloelopinent 
Founder-CEO 
Business and product 
de%'elopinent 
Secured funding 
Fo*inder-CEO 
Business and product 
development 
Secured funding 


B.S. 

computer 
science 


IT co«sultan(/nMn.<grr 
Developed British T\' 
show 


Slandup 
convediji) 


Curti&, 
Doug" 


Lynx 
SolutioiH* 


Founder-CEO 

Business development, hired 

team, secured funding 




Veteran cntrcpreiKur 




Daniekon, 
Ant|c 


Zipcar 


Founder-VP envirnnmental 
affairs and strategy, built 
relationships with car companies, 
tpecilied in<ar technology, 
environmental issues 


Ph.D. 

geochcniistiy 


Supervisor of 
undergraduate cnergy- 
polKy research at 
Harvard University 
Car sales 




Evam, 
Clark' 


Lynx 

Solutions' 


Interiin President/CEO 
Business development, secured 
funding, scaled company 




SVP of consumer 
products company 




GbMrr.Vtill 


Pandora 


CTO 

Tedmical development 


Degrees 
in math, 
computer 
science, and 
physis-s 


Cofounded technology 

startup 

IT consultant 


Amateur 
muskian 


Gh», Noah 


Odeo 


Founder-CEO 
Product development 




Worked on an audio 
application for Biogger 




HouriKan, 
Meg 


Biogger 
toriginally 
Pyra Labs) 


Founder- Vice President 
Business .iiid product 
des'elopinenc, managed 
employees, and provided some 
seed funding 


English 
degree 


Tcchnolog}' consultant 




Kamen. 
Dean 


Segvk-ay 


Founder 

Business aivd product 
developjnent 

Securing funding, hiring tram 


High school 


Inventor of several 
highly siK'cesshil 
medical prodiKts 
Founder-CEO DEKA 




Kendk. 

Candacc 


Kendle 
International 


Founder>CEO 


Phann.D., 

D.Sc. 


Senior faculty positions 
at several leading 
schools and colkrges o( 
pharnucy and medK'ine 




Khullcr. 
Vivck 


Siiuitix 


Fi>under-CEO 
Business and product 
development 

Pitcheii idea to investors and 

VCIUKS 


M.S.. MBA 


5 years at Bell .'\tlaniic 
as a programntcr and 
IT manager 




Knisht, Phil 


Nike 


Founder-President and director 
Business development 
Secured funding 
Negotiated contracts with 
suppliers 

Hired viles and nunagement team 


-MBA 


Accountant 


Track-and-field 
athlete in high 
school and 
college 


Kraft. Jon 


Pandora 


CEO 

Business development 




Eounder-C;EO of 
enterprise database 
startup 




Kraus, Janet 


Circles 


Cofounder-CEO 


MBA 


Technology consulting; 





director of social 
marketing for The Body 
Shop 



Ludvik^M>n. 
Ceorg 


UpD<»wii 


Co-CEO 

B«isine^s development support 
Sales and marketing strategy 
development 


MBA 


Computer 

programming 

Sales 

Cofounded computer 
game startup 




Lunt, ErK^ 


DKA 

Spyonii 
FeedBurner 


Co-VP technology-product 
development 

CTO-priKluct development 
CTO-product development 


B.S.L 

mechanical 

engineering 


Software programmer 
at Andersen Consulting 




McQuillrn. 
Tim* 


SrrongMail 


Cofoundcr 

Chief Information Officer 
Director 


B.A. 

psychology 


Parmer, business 
development at IT 
staffing companies 
Head of IT, sales, data 
centers 




Mciscn- 

hcimcr. 

Mike' 


Ockham 
Technologies 


VP Product Management 
Product development, sales, 
and marketing 
Helped secure funding, 
hired team 


MSM 


Sales and marketing 
consultant 




Milmo. 
James' 


Lynx 

Soluciom* 


Fouiider-President-Chainnan 
BuiiiHrss development, secured 
funding 

Hired team, sales and marketing 


Government 

major 


Campaign worker 
Apprenrice to a toy 
inventor 

Prior failed startup 


Member of 
entrepreneurship 

club 


Minal, 
Saurabh 


Sin,irti\ 


( <>f<Kiiider 
Business and product 
de%'elopment 


MBA 


liiveMim-m Kinkiiig 
internship 




NjILs. Barry 


Masergy 


Founder^CEO 
Business and prodiKt 
development, secured funding, 
hired employees, sales and 
marketing 


MBA 


GTE for 25 years, from 
technician to SVP 




Olcchowski. 
Sieve 


Spyonit 
FtedBurncr 


Programmer 
Businevs development 

COO 




Software programmer 
at Andersen Comultuig 




Pascal, 
Javier* 


Lynx 

Solutions' 


Foiinder-CTO 

Product developmciu, hired team 


Engineering 
degree 


Engineer 




Reich. 
MkIucI 


UpDown 


Founder-C EO 

Business development 

Secure funding, recruit team/sales 

force 


MBA 


Small chainp.igiie 
distributor startup; 
consultant; maruger in 
startup 




Savage, 
George 


Proieus 
Biomedical 


Cofoundei, Chief Medical Officer 

(CMO) 


B.S. 

biomedical 

engineering. 

M.D..MBA 


Cofounded other 
biotcch linns (with 
Andrew Thompson) 




Sherbrooke, 
Karhy 


Circles 


Cofoundec Pretident^COO 


MBA 






Shohc. Matt 


Spyonit 
FecdBunier 


Product design 
Product design 


Graduate 
degree 


User-interface designer 
at Andersen Consulting 




Sfrohtn. 
David* 


Wily 

Technolog)' 


Partner: (Jreylock Partners 
Member of Board of Direcrois, 
adtisor, investor 


MBA 


AiGreylock since 1980 




Thien, 
Genevieve 


Sinerciiy 


Kounder-C:E<.> 
Business and product 
development 

Hired team, sales and marketing 
Secured funding 


B.A. tjiglish 
Graduate 
degree in 
imisic (operal 


Technical writer 
Babysitter 


Opera singer 


Thompson, 
Andrew 


I'rotcui 
Biomedjcdl 


Ciofounder-C.F.O 


MA. 

engineermg 
and 


Cofounded other 
hiotech firms (with 
George Savage I 





ediKation; 
MBA 



TrjLchiman. 
L«« 


Tranwentiv* 
Mfga*ervCT* 


VP Oprraliom, htvt i W 
Rvsrnicnircd salcs/ni.ukcTing/ 
customer service; made several 
suMng changes lu scale 
company 

Enpneered and necoiiared sale of 

company 

CEO 

S«urrd funding; markering, 
sales, and business plan 
de%'elopinent; created RrurKial 
and iiileriul controls 




Direvtor of corporate 

development 

VT Business 

development 

PreMdent/CEO of 

several technology 

startups 




Truong. 
Phuc 


lIpOowii 


Co-CEO 

Product development 


A.B. 

econonucs 


TechnoUqty startup 
Contract IT coiisultinR 




Wtstcrgrcn, 
liin 


Pandora 


Founder-Chief Music Officer 
Music database dcs-clopntent 


B,A. political 
science 


Composer 
Rock hand krader 
Nanny 


Jazz pianist 


Vl'illiams, 
Evan 


Bluggcr 
torifLinally 
Pyra Ijhs) 

Odeo 


Founder-CEO 
ProdiKl and business 
de>-elopinenr, secured funding, 
hired employees 
CofouiKler, advisor, investor 
Evenriully CEO. secured funding, 
hired employees, businesi 
development 


High school 


Self-tauglu Vt'eb 
developer; marketing 
coordinator; developed 
websites for HP and 
Intel 


Two yean 
college 

Prior failed 
startup 


Vt'illiamv. 


Wily 

Technology' 


Pre»ident/CEO 

Hired to replace founder-CEO 
Lew Cirne 


B.S. 

ni.itheniatics 


22 years at IB.VI in 
vano«is leadership role* 
Digital Research - 
President/CEO 
Novell-E\'P Sales 
Illustra InforinatKin 
Technologies - 
President/CEO 
Quokka Sportv 
founding director 




Mark 


Pri>r«t» 
bii>in«dical 


Cofounder, CTO 


B.S. civil 

engineering. 

B.A. 

architecture. 
M.S. 

aeronautics 
and 


Background in micro- 
electrk'al mechanical 
systems (MEMSt 
technology; held se»-eral 
patriKs in atomK and 
iMcmtedical prodiKts 
and processes 





astronautics, 
Ph.D. 
elcctrK'al 
engineering 



^ Disguised name. 

^ John Bohan, "Biography," http://www.johnbohan.com/bio.html (accessed October 2010). 

9 

Exago, Inc., "About Us," httpy/www.exagoinc.com/exago_aboutus.php (accessed October 2010). 

BrightTag, "Management Team," http://www.thebrighttag.com/who_is_brighttag.php (accessed October 2010). 

^ Businessweek.com, "StrongMail Systems, Inc., Executive Profile," 

httpy/investing.businessweek.corn/research/stocks/prrvate/person.asp?personId=9328699&prrvcapId=8594882&previousCapId=8594882 
&previousTitle=StrongMail%20Systems,%20Inc (accessed October 2010). 

^ NewSigma, "Management Team," httpy/www.newsigma.com/Management-Team.html (accessed October 2010). 

^ Greylock Partners, "David Strohm," httpy/www.greylock.com/team/team/14/ (accessed October 2010). 

"7 

Businessweek.com, "IBM Executive Profile," httpy/investing.businessweek.com/research/stocks/people/person.asp? 
personId=234687&ticker=IBM:US&previousCapId=19153&previousTitle=Benchmark%20Capital (accessed October 2010). 



NOTES 



CHAPTER ONE 



1. (Gorman and SahJman, 1989) The authors surveyed 49 well-established venture capitalists (VCs) about 96 of their portfolio 
companies that either had failed or were in danger of failing and asked the VCs to choose up to 3 out of a choice of 11 factors that were 
the major contributors to these companies' difficulties. For 91 of the 96 companies (i.e., 95%), the VCs cited problems within the 
management team as a top-three contributing factor, ranking it the most important contriljuting factor for 61 of those companies (ie., 
65%). 

2. (Kaplan et al, 2004) The authors of this study analyzed 67 internal investment memoranda from 11 VC firms and categorized the 
various analyses of internal, external, and implementation strengths and weaknesses. The internal weaknesses that led the list for 61% of 
the startups included such items as "CEO is a rather difficult person," "Incomplete management team," and "Must strengthen 
management team and ensure involvement of VC as chairman. Will have to hire CEO." 

3. (Bhide, 2000:1) He continues, "Cliche and anecdote — or laments about the ineffable nature of entrepreneurship — dominate the 
discourse about new and fledgling businesses." 

4. These early decisions are particularly formative for the organization; in the earfy days of a startup, founders have much more 
control than they do after inertial forces set in (Hannan et al, 1989) and after the effects of that early imprinting begin constraining the 
organization (Boeker, 1988, 1989). 

5. (Stevenson et al, 1990:23) This definition will play a central role in our analyses of the core trade-offs faced by founders. 

6. (Aldrich et al, 2006) 

7. (Baker et al, 2003; Baker et al, 2005) 

8. When almost 3,000 small business owners were asked about the odds of success for their own businesses, they ranked their own 
chances as being an average of 8.1 out of 10. When asked about the odds of success for businesses similar to their own, they ranked 
them as being 5.9 out of 10. For more details, see Cooper et al (1988). 

9. Busenitz et al (1997) surveyed 176 founders of firms in the plastics, electronics, and instruments industries and 95 managers from 
large public companies. They asked six fact-based questions about death rates from diseases and questions that gauged the confidence 
the respondent had in his or her answers. The founders were overconfident in their choices in five out of six categories, whereas 
managers were overconfident in only three out of six categories. The researchers' overconfidence variable predicted sorting into 
entrepreneurship 70% of the time. 

10. Hayward et al (2006) argue that as overconfidence increases, founders attract smaller initial resource endowments because 
they overestimate customer demand, neglect the reference group of competitors, underestimate requirements for legitimacy, and 
overestimate their ability to develop social ties and secure needed resources. 

11. In a sample of 201 founder-CEOs taken from the Dun & Bradstreet database, each respondent was given a six-item Life 
Orientation Test that measures a general sense of optimism about life. This variable was associated with a 20% decrease in revenue 
growth and a 25% decrease in employment growth. The authors suggest that optimism can cause startups to fail because entrepreneurs 
hold unrealistic expectation, discount public information, and mentally reconstruct past experiences to avoid contradictions. The negative 
relationship between optimism and performance increased in the presence of environmental dynamism (Hmieleski et al, 2009). 

12. (Baron, 1998) 

13. (Emerson, 1962; Pfeffer et al., 1978) 

14. (Stevenson et al, 1990) 

15. (Amit et al, 2000; Carland et al, 1984; Sapienza et al, 2003) 

16. (Wadhwaetal, 2009:13) 

17. (Hurst et al., 2010b, table 7) 

18. (Scitovszky, 1943:57) 

19. (Kirzner, 1973; Schumpeter, 1942) 

20. (Amit etal, 2000:120) 

2 1 . (Moskowitz etal, 2002) 

22. (HaU etal, 2010) 

23. (Hamilton, 2000) 

24. (Hamilton, 2000:622-623) Although this study has recently been criticized for relying on data sources that understate 
entrepreneurs' earnings by a significant percentage (Hurst et al., 2010a), even after adjusting for that underreporting, the result still holds 
that entrepreneurs earn less than would be expected, though the difference is smaller Amid multiple guesses as to why this might be 
true, Hamilton speculates that the reason might be "that many entrepreneurs receive substantial non-pecuniary benefits, such as 'being 
their own boss.' " This possibility is one we will examine closely. 

25. Two of these authors (Moskowitz et al, 2002) even asked about this "puzzle" in the title of their study. 

26. (Zalezniketal., 1975) 

27. (Boeker et al., 2002; Wasserman, 2003) 

28. (Wasserman, 2006b) 

29. (Mills, 1956:162) 



30. Renaissance Capital's data on IPOs from 2000 to 2009 show that technology startups accounted for almost 30% of all IPOs, and 
life sciences and healthcare for 18%. The financial industry accounted for almost 12%, and no other industry was above 8.3%. 

31. According to my analyses of annual reports from the Center for Venture Research at the University of New Hampshire, which 
compiles perhaps the most reliable data on angel investments, technology investments (including software, hardware, telecommunications, 
and IT services) accounted for 45% of angel investment across the decade and life sciences investments (biotechnology, Hfe sciences, 
and health) accounted for an additional 29% (Sohl, 2001-2009). 

32. According to Thomson's Venture Expert Database, accessed September 16, 2010, technology investments (including Internet, 
computer software and services, communications and media, semiconductors and other electronics, and computer hardware) accounted 
for 56% of VC investments from 2000 to 2009 and life sciences investments (medical/health and biotechnology) accounted for 15%. 

33. (Aldrich et aL, 2006; Hurst et aL, 2010b) 



CHAPTER TWO 



1. (Higgins, 2005a) 

2. (Higgins etal., 2010) 

3. This particular analysis included 2,732 first-time founders. Similarly, in a recent study of 86,000 science and engineering graduates, 
age was not a significant predictor of the transition to self-employment. Researchers drew their data from the National Science 
Foundation's Scientists and Engineers Statistical Data System (SESTAT) and focused on data from 1995 to 2001. Every individual in the 
sample had at least a bachelor's degree in a science or engineering field (Elfenbein et al, 2009). 

4. This is consistent with other research, such as a study of 1,547 small-business entrepreneurs in the United States that found that 
the average age of the entrepreneurs was 36.7 years old, with a standard of deviation of 9.4 years. The sample was drawn from the 
National Federation of Independent Businesses (Gimeno et al., 1997). 

5. Taking a multi-decade view of a subset of founders, the age distribution of founders appears to have become younger over the 
past few decades. A survey of MIT graduates found that between the 1950s and 1990s, founding ages dropped 12.5 years, leading the 
authors of that study to conclude that entrepreneurship was transitioning from a "mid-life career change" to an earlier-in-career option 
(Hsu et al.,2007). 

6. In a sample of 229 founder-CEOs of architectural woodworking firms, passion influenced motivation by affecting the founder's 
confidence in his or her skills, the founder's goals for sales and employment growth, and how that growth orientation affected the firm's 
vision statement (Baum et al, 2004). 

7. For example, Busenitz et al (1997) compared the risk-taking preferences of 176 founders of firms in the plastics, electronics, and 
instruments industries to those of 95 managers from large public companies. Risk-taking preferences were measured using the Jackson 
Personality Inventory and did not significantly predict entrepreneurship. A meta-analysis (Stewart et al, 2001) of prior studies of risk 
differences between entrepreneurs and managers seems to suggest support for the proposition that entrepreneurs are more risk-tolerant 
than managers, especially among growth-oriented entrepreneurs. However, the study's methodology has been criticized for treating as 
independent multiple studies that used the same underlying dataset of growth-oriented entrepreneurs, and the authors have been criticized 
for not being "conservative" enough in interpreting their results. (See Miner et al [2004] for details.) More broadly, decades ago, 
academics made a concerted effort to nail down the core characteristics of the "entrepreneurial personality," but in vain. "In the end," as 
one expert summed it up, "it turned out that there is no such thing as a well-defined 'entrepreneurial personality.' " 

8. (Higgins, 2005a; Higgins et al., 2010) 

9. In a 2007 study of 20,000 individuals from 27 European countries and the United States, Stam et al. (2008) found that respondents 
with self-employed parents were more likely to be entrepreneurs and their startups were less likely to fail The authors attribute this to 
indirect learning from observing the entrepreneurial actions of role models. 

10. The survey uses dozens of scenarios to assess how important each motivation is to the person. 

11. The database had complete work-reward surveys for 27,357 people, including 1,810 male entrepreneurs, 13,939 male non- 
entrepreneurs, 404 female entrepreneurs, and 11,204 female non-entrepreneurs. 

12. See Gimeno et al (1997) for more details on the differences between formal and tacit human capital, and Aldrich et al (2006, 
chap. 4) for a discussion of procedural memory, declarative memory, and the accumulation of the tacit knowledge of how and when to 
invoke each type of memory. 

13. Building on Fiske et al (1991), Aldrich et al (2006:76) define schemata as "cognitive structures that represent organized 
knowledge about a given concept or type of stimulus." They define three levels of schemata: person schemata, which affect how people 
assess new people they encounter, the personality traits they associate with them, and the groups of personality traits they associate with 
each other; role schemata, which affect the behaviors they expect of people in particular social positions (and which they expect of 
themselves when they are in those positions); and event schemata, which affect the sequences of events people expect in situations 
similar to ones they have experienced. 

14. Other factors shaping people's schemata include advice from mentors and experts and the structure and actions of other 
organizations they have read or heard about (Aldrich et al, 2006). 

15. (Beckes-Ge liner et al, 2008) The authors use a sample of 2000 German university students from Cologne. "Willingness to 
become an entrepreneur" was measured on a 4-point scale in response to a single item on the questionnaire. Of the sample, 16% was 
identified as willing to become an entrepreneur "Broad human capital portfolio" was measured as a count of the different types of work 
experiences and academic degrees an individual holds. This variable was associated with a 116% increase in willingness to become an 
entrepreneur. 

16. (Shane, 2008:47, italics in original) Likewise, Davidsson et al. (2003) matched a group of 623 nascent entrepreneurs from 
Sweden with a group of 608 non-entrepreneurs. "Explicit human capital" was measured by the highest level of formal education 
completed. This variable was associated with a 16.7% increase in the likelihood of being sorted into the nascent entrepreneur group. 

17. Across all of the startups in my dataset, 25% of founder-CEOs had pre-founding management experience: 27% for technology 
founder-CEOs and 19% for life sciences founder-CEOs. 

18. Meyer et al (1990) undertake a close study of 25 founder-CEOs of high-tech firms in Boulder, Colorado. They find that 
founders are predominantly technical in their skills and identify four aspects of this mindset that make technical founders averse to 



developing management skills: Technical CEOs (a) are detail-oriented to the exclusion of the big picture, (b) lack flexibility, (c) dislike 
formal, structured environments, and (d) are "elitist" and unwilling to seek expert management advice. 

19. (Bhide, 2000:47) 

20. In their groundbreaking theoretical study on the effects of an organization's top managers, Hambrick et al. (1984) propose that 
functional backgrounds often have an important influence on firm strategy. 

21. (Shane 2008:38) Similarly, in a study of 100 companies from the Inc. 500 list of the fastest-growing private companies in the 
United States, 40% of the founders had no prior experience in the industry in which they founded their startups and many others had only 
shallow or narrow experience in the industry (Bhide, 2000). 

22. Indeed, having a well-developed cognitive structure regarding a context can prevent a person from seeing disconfirming or novel 
evidence and can inhibit creative problem solving (Walsh, 1995). 

23. Examining 1,849 founders who registered a new business in Germany between 1985 and 1986, Bruderl et al. (1992:237) found 
that industry-specific experience was associated with a 33.2% decrease in the rate of startup failure. Industry-specific experience was 
significantly and positively related to seed capital, legal incorporation, and employment growth, leading the authors to conclude that 
"starting a business without experience in the industry sharply increases the mortality rate." 

24. (Higgins, 2005b) 

25. Previous startup experience was associated with a 78% increase in the likelihood of being in the nascent entrepreneur group 
(Davidssonet aL,2003). 

26. For more details, see Baron et al. (2006). People with more entrepreneurial experience evaluated opportunities using the 
following dimensions: solving a customer's problems, generating positive cash flow, manageable risk, speed of revenue generation, and 
the existence of others in their network with whom to develop the venture. People with little or no entrepreneurial experience used the 
following dimensions: how novel the idea was, whether it was based on new technology, superiority of product/service, potential to 
change the industry, and intuition or gut feel 

27. (Shane et aL, 2000) 

28. (Gompers etal.,2005) 

29. In this study (Elfenbein et al., 2009), companies with 100 or fewer employees had fewer than 20% of all employees in the 
sample, but produced 64% of the founders. The researchers drew their data from the National Science Foundation's SESTAT and 
focused on data from 1995 to 2001. It should be noted that every individual in the sample had at least a bachelor's degree in a science or 
engineering field, and that the researchers included in their sample only men who worked full-time. 

30. Using a sample of 2,692 MBA graduates of a U.S. business school, Dobrev et al. (2005) found that the size of the organization in 
each person-year is associated with an 8% decrease in likelihood of transitioning to entrepreneurship in the next person-year. 

31. (Higgins, 2005b) 

32. For broad coverage of the types and uses of social capital, see Kim et al. (2005). 

33. Across a 197I-198I sample of 1,443 white men, including 89 entrepreneurs, the average probability of being a financially 
constrained entrepreneur was 3.75%, while the probability of being an unconstrained entrepreneur was 0.06% (Evans et al., 1989). 

34. Blanchflower et aL (1998) used a sample from the National Child Development Study, a longitudinal birth-cohort study including 
all individuals in the United Kingdom born between March 3 and March 9, 1958. Data for the study were collected in I98I (age 23) and 
1991 (age 33). 

35. In a sample of 275 high-potential startups that went public in 1996, for all levels of human resources (composite of industry 
experience, startup experience, and VC directors), social resources (composite of business network, personal network, and underwriter 
prestige) were associated with a 23% increase in financial capital (Florin et al., 2003). 

36. Being involved in a business assistance network and having family and friends encourage the venture were associated with a 
faster pace of completing certain "gestation activities" (Davidsson et al., 2003). 

37. This means that someone who started working at age 21 founded a startup 25 years later at age 46. 

38. Bruderl et aL (1992) found a nonlinear relationship between years of work experience and rate of startup failure. The 
relationship was negative until 25 years of experience, after which it became positive. 

39. In a sample of 5,000 respondents who were graduates of Stanford's MBA program, the likelihood that an employment spell is 
entrepreneurial is positively and strongly associated with the number of prior roles held by an individual at all of his or her past jobs 
combined (Lazear, 2004). 

40. Interestingly, Elfenbein et aL (2009) suggest that having children in the household may be associated with an increased chance 
of making the leap into founderhood, a result that should be validated and whose underlying reasons should be explored. It is possible that, 
for potential founders, the arrival of an infant differs from having adolescent or teenage children in the household, or that having an initial 
child differs from already being an experienced parent. 

41. This survey was of 549 founders from a wide variety of industries, including computer and electronics, healthcare, aerospace and 
defense, and services (Wadhwa et aL, 2009). 

42. Having an unemployed spouse had a small but significant negative effect on transition to entrepreneurship (Elfenbein et aL, 
2009). 

43. In a study of 5,112 startups that had received VC funding from 1986 to 1999, Gompers et aL (2005) focused on the public 
companies that spawned spin-off startups and found that an increase in sales growth from the 25th percentile to the 75th percentile 



decreased spawning rates by 4.6%. 

44. Elfenbein et al. (2009) used paid wage as a proxy for performance. Individuals from the 5th percentile of the wage distribution 
were 35% more likely than individuals at the median to enter into entrepreneurship, while individuals at the 95th percentile were 28% 
more likely. 

45. Respondents who had received £5,000 or more were twice as likely to become entrepreneurs than those who had received 
nothing (Blanc hflower et aL, 1998). 

46. (Hart et al., 2003) 

47. See Bhide (1994). A separate study of small businesses found that 56% of founders identified their new business ideas based on 
prior experience in the market or industry (Hills et al., 2004). 

48. For the most influential of these, see Porter (1980). 

49. (Bruderletal., 1992) 

50. Christensen et al. (1996) constructed a complete census of products and specifications for every model introduced in the 
computer disk-drive industry from 1975 to 1990. Based on interviews with 70 executives in 21 companies, they created a model of 
"sustaining" versus "disruptive" technological change. (Other researchers refer to these differences as "competence destroying" and 
"competence sustaining.") Aldrich et aL (2006) suggest that established firms succeeded in introducing sustaining technologies but failed 
to capitalize on disruptive technologies because their marketing focus groups could not find a use for them. 

51. (Hannanetal., 1989) 

52. Pennings (1982) compared data on new organizations in three industries with census data on major American cities. Cities with 
increased organizational birth rates had larger industry populations, residents who had more income and more savings to invest in 
business, above-average domestic immigration, and local colleges and universities. 

53. Respondents' subjective assessment of the competitive intensity of the environment facing them was associated with a 20% 
decrease in organizational survival rates (Bruderl et aL, 1992). 

54. For an overview of network effects, including the sources of network effects and the characteristics of industries affected by 
them, see Economides (1996). For an examination of the two-sided markets in which network effects can be particularly strong, see 
Eisenmann et aL (2006). 

55. (HamermeshetaL, 2008:3) 

56. Cooper et aL (1988) surveyed 2,994 owner-managers (including founders and nonfounder-CEOs) in 1984 and 1985. The sample 
was drawn from the 1985 National Federation of Independent Businesses listing, covering all industries and all geographic regions of the 
United States. 

57. The survey included six questions about death rates from diseases in the United States, along with follow-up questions that tested 
the confidence the respondent had in his or her answers. Overconfidence was measured by taking into account the mean confidence 
level respondents had for all their answers, by subtracting the mean from the actual percentage of correct answers (Busenitz et aL, 
1997). 

58. Bernardo et aL (2001) constructed a theoretical model of social populations in which entrepreneurs are individuals who will be 
more likely to act on their own private information as opposed to public information. As overconfidence increases, founders are more 
likely to act based on their own information. 

59. In a study of 143 founders of high-growth printing and graphics firms, Baum et aL (2010) found that respondents' answers to 
questions about their confidence in their ability to perform business tasks was associated with a 33% increase in "swift action" in 
problem-solving scenarios. 

60. Lovallo et aL (2003) outline the "planning fallacy" to which founders are susceptible because business plans are anchored in 
proposals meant to market the organization to resource providers and because entrepreneurs tend to focus on internal causes instead of 
external competitors. They recommend that entrepreneurs creating business plans should minimize the risk of succumbing to the planning 
fallacy by focusing on a reference group. 

61. Hayward et al (2006) propose that overconfident founders may attain smaller initial resource endowments and may be more 
committed to the initial opportunity, causing them to be less flexible and more prone to failure. 

62. (Hmieleski et al, 2009; Malach-Pines et aL, 2002) 

63. Hmieleski et aL (2009) surveyed 201 founder-CEOs of U.S. businesses operating in 114 different industries. 

64. (Kawasaki, 2006) 

65. (Blank, 2009) 

66. (Wasserman, 2008a) 



CHAPTER THREE 



1. (Stinchcombe, 1965) 

2. These analyses used data on 1,531 core founders from my 2006-2009 CompStudy datasets. The effect of having prior 
management experience had a strongly significant impact on the decision to solo-found. The effects of (a) having been a serial 
entrepreneur and (b) having more years of prior work experience were much weaker, possibly because serial entrepreneurs and people 
with more work experience may have counterbalancing factors that lead them to cofound more often. For instance, they may have 
developed better networks with which to find stronger potential cofounders, making it more attractive to cofound. 

3. (Awadetal, 2010:11) 

4. In a study of 1,849 founders of German small businesses, Bruderl et al (1992) found that having prior experience in the industry in 
which the business was founded (as did Barry NaUs) was associated with a 33.2% lower chance of business failure. 

5. The Panel Study of Entrepreneurial Dynamics (PSED) is a research program, headquartered at the University of Michigan, 
consisting of two longitudinal projects designed to identify and interview "nascent entrepreneurs" — individuals involved in the early 
process of creating firms. PSED 1 began in 1998 with a cohort of 830 nascent entrepreneurs who were interviewed every 13-14 months 
in 1998-2000. PSED 11 was a revised and enhanced version of PSED 1 in which a cohort of 1,214 nascent entrepreneurs were 
interviewed every 12 months beginning in 2005. 

6. It is interesting to note that, within the PSED small-business dataset, there are also some gender and ethnic effects on solo 
founderhood. Ruef et al (2003) estimate that women are 1.2 times more likely than men to solo-found, which is consistent with the data 
presented in Chapter 1 on the gender differences in control motivations. Also, minority solo-founding is 1.2 times more likely than solo- 
founding by nonminorities. 

7. (Stinchcombe, 1965) 

8. (Hackman et al., 2000; Mosakowski, 1998) 

9. (Eisenmann et al., 2006) 

10. Among teams in the computer industry, Haleblian et al (1993) found that larger teams were associated with a 49% increase in 
performance. They conclude that in turbulent environments, the information-processing benefits of a larger team outweigh the added 
coordination and communication costs. 

11. (Aldrich et al., 2006; Eisenhardt et al., 1988; Keck, 1997; Virany et al, 1992) 

12. (Haleblian etal, 1993) 

13. (Bruderl et al., 1992; Eisenhardt et al., 1990; Singh et al, 1990) These findings have interesting implications for Stinchcombe's 
(1965) theory about the internal causes of the liability of newness, suggesting benefits of founding teams that may counterbalance the 
risks he highlights. 

14. (Reynolds et al, 2008; Shane, 2008) Only 13% of PSED small businesses were founded by more than two people (Aldrich et al., 
2004). In the Inc. 500 list of the fastest-growing private companies in the United States, approximately 30% were started by solo 
founders (Bhide, 2000). 

15. As we will see in subsequent chapters, such "founders" may also own significantly less of the startup's equity than their 
cofounders, or have less-central roles. 

16. (Bussgang, 2010) 



CHAPTER FOUR 



1. (RuefetaL,2003) 

2. Baker et al. (2003) present a theoretical framework of founder improvisation or "bricolage," in which time -constrained 
entrepreneurs use the most readily available resources to construct their startups. 

3. Aldrich et aL (2006) argue that hiring more strangers increases the difficulty of creating organizational boundaries between 
members and norunembers, and propose that, in new organizations, hiring more strangers should be associated with higher turnover. 

4. (Hambricketal., 1984) 

5. (Amason, 1996; Knight et al., 1999) 

6. In a 1979 sample of 79 individuals from 20 work groups in franchises of a national chain, more heterogeneity in tenure at the 
organization and in age was associated with 54% and 25% decreases, respectively, in group- level social integration measures (O'Reilly 
et al., 1989). 

7. For instance, teams with diverse experiences, such as having created comic books in a broader range of genres, are better suited 
to producing innovative products, though extremely high levels of diversity can be counterproductive (Taylor et al., 2006). In the 
founding-team context, see also Ucbasaran et aL (2003). 

8. In turbulent contexts, functional heterogeneity is associated with increased returns (Keck, 1997; Wiersema et aL, 1992). 

9. (Beckman, 2006) Each additional diverse-employer relationship meant the firm was 1.2 times more likely to tend toward 
exploration; each additional common-employer relationship meant the firm was 1.4 times more likely to tend toward exploitation. 

10. (Kim et aL, 2005) The authors argue that founding teams with non-overlapping networks gain more rapid access to diverse 
information and opportunities. Similarly, homogeneous teams tend to have less collective social capital because the members' networks 
tend to overlap (Burt, 1992). The author presents a theoretical model of social capital that proposes that networks rich in structural holes 
(Le., communication nodes that are not in contact with each other) should give the focal actor advantages from nonredundant 
information. 

11. Beckes-Ge liner et aL (2008) used the Cologne Founder Study, a dataset of German university students from five universities in 
metropolitan Cologne in 1999 and 2000, to estimate willingness to become an entrepreneur Similarly, individuals with more ties to diverse 
contacts in nonoverlapping networks (indicated by involvement in multiple business networks such as trade associations and rotary clubs, 
seeking advice from small-business assistance organizations, etc.) begin new companies more often (Davidsson et aL, 2003). In the latter 
study, which compared 623 nascent entrepreneurs to 608 non-entrepreneurs, membership in multiple business networks was positively 
and significantly associated with nascent entrepreneurship. 

12. Interestingly, homophilic tendencies may be exacerbated by student self-selection into schools and by school admission 
committees, both of which may serve as a strong homogenizing influence on classmates. 

13. (Wozniak, 2006:279) 

14. (Wozniak, 2006:12, 43, 147) 

15. (SchaubroecketaL, 1998) 

16. Genesis 2:18, Artscroll translation of the interpretation of Rabbi Shlomo Yitzchaki (aka Rashi). 

17. (RuefetaL,2003) 

18. (Wasserman et aL, 2008) 

19. In Baron et aL's (1996) sample of 172 high-technology startups in Silicon Valley, the commitment blueprint (for details, see 
Chapter 8) was adopted by 83% of firms in which famify or friends of the founder were listed as key partners, compared to just 30% of 
the total sample. 

20. (Eisenhardt et aL, 1990; O'Reilly et aL, 1989; Pennings et aL, 2010; Ucbasaran et aL, 2003) In the context of Wall Street 
analysts, teams of prior coworkers who moved together to a new firm performed better than analysts who moved alone (Groysberg, 
2010). 

21. (Wozniak, 2006:172) 

22. The fear of conflict within the team causes artificial harmony (Lencioni, 2002). 

23. (Wozniak, 2006:147) 



CHAPTER FIVE 



1. (Roberts etaL, 2009) 

2. A notable exception is professional-services firms, where small firms tend to be structured as "upside-down pyramids" until they 
grow older, bigger, and more formalized, at which point they become much more pyramidal (Wasserman, 2005). Another exception, 
closer to home, is the early-stage venture capital firms that invest in startups. For details about the performance implications for VC firms 
of adopting pyramidal versus upside-down pyramidal structures, see Wasserman (2008b). 

3. These analyses included 952 idea and non-idea founders from technology and life sciences ventures surveyed in 2008-2009. 

4. (Jehn, 1997) 

5. In Weberian sociology, ideal types are formed from characteristics and elements of a phenomenon. Ideal types focus on elements 
found in many cases of the phenomenon, but may not correspond to all of the elements of any particular case. They enable the creation 
of "unified analytical constructs" for analyzing those cases and facilitate comparisons across cases (Weber, 1949/1997). 

6. Weber provided the classic analysis of the value — and challenges — of hierarchy. His core example is the bureaucratic 
organization, which emphasizes technical expertise and the division of labor among the offices in a hierarchy. Such an organization has 
far greater efficiency, reliability, and stability than other forms of organization (Weber, 1946). 

7. In a study by Haleblian et aL (1993), firm performance was measured as an aggregate of return on assets, return on sales, and 
return on equity. CEO dominance — an estimate of how unevenly power is concentrated in the company's CEO — ^was measured by 
averaging the coefficients of variation for each of 10 indicators of power — including compensation, titles, board membership, education, 
and expertise — within the top management team of each company. 

8. Eisenhardt et al. (1988) focused on high-technology firms in "high-velocity" environments. 

9. In a study of 26 large firms in the computer industry, CEO dominance was associated with a 19% decrease in firm performance, 
measured as an aggregate of return on assets, return on sales, and return on equity (Haleblian et al., 1993). 

10. An in-depth study of eight teams in the microcomputer industry found that consensus-based approaches took too much time and 
caused slow decision making, leading firms to miss opportunities in the fast-moving industry (Eisenhardt, 1989). 

11. (Rogers et al., 2006:56) 

12. Panel at meeting of The Indus Entrepreneurs-Boston (TiE-Boston) group, November 2005, moderated by Noam Wasserman. 

13. Venture capitalists also tend to disapprove of egalitarian teams, preferring instead to have a clear "buck stops here" CEO with 
whom they can interact. This antagonism toward egalitarianism in startups is particularly interesting because most early-stage VC firms 
are also run as egalitarian partnerships, with multiple general partners making decisions collectively (Wasserman, 2002, 2008b). 

14. Blau (1970) showed that an increase in organizational size is usually accompanied by an increase in structural differentiation, 
resulting in greater heterogeneity of the work performed across the organization. This increased heterogeneity deepens the division of 
labor and introduces problems of coordination and integration. 

15. (Hellmanetal.,2002) 

16. Eisenhardt (1989) examined this approach in her study and proposed that it would be an effective model for making fast 
decisions in high- velocity environments while still achieving universal buy-in. 

17. In an in-depth study of strategic decisions from 45 teams in the food-processing industry, team size was associated with a 13.9% 
increase in affective conflict (Amason et al., 1997). 



CHAPTER SIX 



1. See Hall et aL (2010) for evidence of this in venture -backed startups, and Hamilton (2000) and Moskowitz et al. (2002) for 
evidence of this in small businesses. Cash and credit constraints can also force founders to accept lower pay or to take more deferred 
compensation than in alternate employment options. 

2. (Northcraftetal., 1987) 

3. For more on free riding and the double-sided moral hazards that can arise from "any type of joint production arrangement where 
two parties each contribute some unmarketed input to the production process," see Bhattacharyya et aL (1995:767). 

4. (Young etaL, 2004:34) 

5. See Appendix B for the specific questions used. 

6. The dependent variable in these ordinary least squares (OLS) models was the percentage of equity received by each founder 
The core independent variable was a dummy variable that captured whether each founder had been an idea person. The control 
variables included controls for the founders' backgrounds (e.g., years of prior work experience, experience in the same industry, whether 
a founder was a serial entrepreneur), early contributions to the startup (e.g., capital contributed), initial position within the founding team, 
the number of founders, the startup's industry segment, and other variables. 

7. The specific size of the idea premium varied slightly by industry and by the type of regression model, but always fell within the 
10-15 percentage -point range. Variations within this range may also be due to the difference in ideas: WeU-developed ideas that are new 
in a fundamental way should receive more of a premium than nascent ideas that are less groundbreaking. This estimate of the idea 
premium was also confirmed by my deeper analyses with Thomas Hellmann of my 2008-2009 dataset. 

8. In the PSED dataset of small businesses, a cofounder 's contribution of seed capital was associated with a 20.6% increase in that 
cofounder's equity stake (Ruef, 2009). 

9. A study of 623 nascent entrepreneurs and a control group of 608 non-entrepreneurs found that previous startup experience was 
associated with a 77% increase in probability of being in the nascent entrepreneur group (Davidsson et al., 2003). 

10. In the Panel Study of Entrepreneurial Dynamics dataset, differential seed capital contribution is associated with a 20.6% increase 
in founder equity stake. 

11. See Wasserman et al. (2010a) for details. The analyses in this paper focused on the results from my 2008 and 2009 surveys, 
which included 1,476 founders from 511 private technology and life sciences ventures. 

12. In these probit models, the dependent variable was a binary indicator of whether the split was equal or not. The models included 
the team's prior relationships, the team's average amount of these four variables, the team's coefficient of variation for each of the four 
team-level variables, founding-team size, and dummies for industry, location, and year of split. 

13. Some types of equality are more equal than others. Our analysis found that the more similar the founders' initial capital 
contributions, the more likely the team was to split the equity equally. In fact, when we analyzed the different types of splits, of all of the 
factors in our regression analyses, the capital-contribution variables almost always had the strongest levels of statistical significance. 

14. In these analyses, the dependent variable was the log transformation of the venture's pre-money valuation during its first round 
of outside financing. The core independent variable was the binary indicator of whether the team had split the equity equally, with an 
accompanying binary indicator for whether the equity split was quick or slow. Control variables included the same variables used in the 
equal-split probit analyses described above: the team's prior relationships, the team's average amount of each of the four background and 
early-contribution variables (prior founding experience, prior years of overall work experience, idea person, and seed capital contributed), 
the team's coefficient of variation for each of the four team-level variables, founding-team size, and dummies for industry, location, and 
year in which the round was raised. Given that some of our ventures had not raised an outside round of financing, = 298 ventures for 
these OLS regression models. 

15. (Bagleyetal.,2003) 

16. (Strasser etaL, I99I) 

17. Each state has its own governing body of law that will determine whether a contract existed, what its terms were, whether a 
breach occurred, and what remedies are available. However, the overall elements of a contract are (a) an offer, (b) acceptance, and (c) 
consideration. If these three elements exist, a contract can be found to have been established under common law. At the same time, oral 
agreements are hard to enforce in situations in which negotiations have been drawn out or are complicated and ambiguous. Individual 
states have adopted statute -of-fraud laws that require certain types of contracts (including, interestingly, marital prenuptial agreements) 
to be put in writing. To avoid the possibility of having an agreement ruled unenforceable, parties should put in writing any contract that 
might take more than a year to perform (Bagley et aL, 2003). 

18. (Malhotra, 2009) 

19. (Shane, 2008:69) In the high-tech realm, the SPEC study found that 15% of its startups had made major changes to their 
business strategies and a much higher percentage had made lesser but important strategic changes (Hannan et aL, 1996). Scholars 
predict that strategic change in new ventures will become even more common as globalization increases (McDougall et aL, 1996). 

20. (Video accompanying Hart et aL, 2003) 

21. (Hegedus et aL, 2001) 

22. Press briefing by former U.S. Secretary of Defense Donald Rumsfeld on February 12, 2002. 



23. (Coleman, 1988) The author argues that dense and redundant networks enforce mutual obligations and thus form the basis of 
social capital. 

24. In "closed" teams or partnerships, free riding and other moral hazards can be solved by separating ownership from control and 
bringing in a principal to monitor the agents' inputs and behaviors. When the venture matures and takes in outside capital, such a shift 
eventually occurs, as discussed in Chapter 9. For more information, see Alchian et al. (1972) or Holmstrom (1982). 

25. To get these descriptive data, I combined my annual cross-sectional datasets from the 2005-2009 surveys in both technology and 
life sciences and focused on the startups that had two or more cofounders still working. The resulting dataset included 2,815 cofounders 
from 1,148 startups. Initial analyses of potential survival biases indicated that it was not a problem, but subsequent research may want to 
take a longitudinal and multivariate approach to such anafyses and assess whether the results are affected by differential survival rates 
within two-founder versus three-founder teams and other such differences. 

26. (Jensen et aL, 1976) 



CHAPTER SEVEN 



1. This is based on the SPEC sample of 172 high-tech firms from Silicon Valley (Baron et aL, 1996). 

2. (Ruef,2009) 

3. (Adams, 1965; Deutsch, 1975; Leventhal, 1976) 

4. (Deutsch, 1975). 

5. (Austin, 1980) Similarly, Kabanoff (1991) suggested that equality rules are consistent with social logics while equity rules are 
consistent with task logics. 

6. (Deutsch, 1975; Leventhal, 1976; Leventhal et al., 1980) 

7. (WassermanetaL,2008) 

8. (Lawler, 1971; Leventhal, 1976) 

9. (Leventhal, 1976) 

10. (Wasserman, 2006b) 

11. (SabherwaletaL,2001) 



PART m: INTRODUCTION 

1. This description of organizational evolution was heavily influenced and informed by multiple frameworks (e.g., Baron et aL, 1996; 
Baron et aL, 2001; Galbraith, 1982; Greiner, 1972). 

2. This is the central message of exchange theory and resource-dependence theory. For more details on these theoretical 
foundations, see Blau et aL (1962), Emerson (1962), and Pfeffer et aL (1978). 



CHAPTER EIGHT 



1. Beginning in 1994, the SPEC study tracked 172 high-technology startups in Silicon Valley Researchers conducted interviews with 
founders, CEOs, and HR directors and supplemented this information with public and private data on strategy, HR practices, business 
partners, financing, and other aspects of the business. 

2. Of the sample, 50.9% did not change their blueprint at all, while 29.7% changed only along one dimension (Baron et al, 2001; 
Baronet al,2002). 

3. Baker et al. (2003) present a theoretical framework of founder improvisation or "bricolage," in which time-constrained 
entrepreneurs use the most readily available resources to construct their startups. 

4. These data include 1,522 nonfounding executives from founder-CEO-led startups in the 2007, 2008, and 2009 annual surveys in 
technology and life sciences. More specifically, the data include 285 CFOs, 202 CTOs and CSOs, 121 COOs, and 914 VPs (heads of 
sales, marketing, business development, engineering, human resources, and — where relevant — ^professional services, 
manufacturing/operations, clinical research, and regulatory affairs). 

5. (Aldrich et al, 2006; Waldinger et al., 2003) 

6. Hsu's (2007) sample included 149 startups participating in a semester-long educational program at MIT known as the 
"entrepreneurship laboratory." 

7. Use of a strong tie was associated with a 40% decrease in occupational status from first job and a 29% decrease in status of last 
job. The sample included 399 males, 20 to 64 years of age, in the tricity capital area of New York State in 1975 (Lin et al., 1981). 

8. (Aldrich et al, 2006) 

9. (Bhide, 2000:49) 

10. (Sherer, 1995) 

11. (Hitt et al, 2001; Wasserman, 2008b) 

12. In an in-depth study of 14 startups in North Carolina's Research Triangle, Baker et al (1994) constructed a punctuated- 
equilibrium model whereby organizations become more formalized over time as founders reach outside of their personal networks for 
more and more specialized employees. 

13. In their analysis of the SPEC sample of high-tech startups. Burton et al (2007) find that position successors (i.e., the second 
generation or more of employees at a company) with different functional backgrounds than the position creator have an 11% higher 
turnover rate. Successors who follow "atypical" position creators (defined as individuals with functional backgrounds that differ from the 
average functional background for their position in the sample) have 15% to 23% higher turnover rates. 

14. (Baker et aL, 1994) Early employees tended to be hired through informal or personal channels, tended to end up in the most 
senior jobs, and tended to be generalists with substantial experience who were willing to accept relatively undefined positions. 

15. (Baker et al, 1994) In the early stages of startup evolution, senior-level employees were mainly differentiated along the lines of 
operational and technical functions; it was only after the depletion of seed capital that clear roles developed for senior marketing and 
sales executives. 

16. (Baron et al, 2001) 

17. (Hart et aL, 2003:9) 

18. (Baker et aL, 1994) In contrast, startups that take on junior employees with a background in large companies may feel pressure 
to provide some of the same human resource management practices to which the new hires grew accustomed at their old jobs. 

19. In early-stage startups, when roles are most likely to overlap, it is harder to measure individual performance and thus to tie 
compensation effectively to individual performance (Aldrich et al, 2006), possibly increasing the emphasis in startups on the equity side 
of the reward structure. 

20. Carpenter et al (2002) found, in their study of a five-year survey of executives conducted by a major compensation consulting 
company in the early to mid-1980s (which included 34,500 executive-year observations from 90 public companies), that an executive's 
total compensation was significantly related to the similarity between the executive's function and the most critical strategic contingency 
facing the company. 

21. In a diverse sample of 209 California companies, franchises, and subsidiaries. Baron et al (1986) found evidence supporting the 
existence of a gender gap in compensation. For an examination of pre -hiring rather than post-hiring effects on gender segregation, see 
Fernandez et al (2005). 

22. The analyses described here are based on my 2008 data, when I added a gender question to my annual survey. The resulting 
dataset included 2,202 C-level and VP-level executives from 459 private technology and life sciences startups. The regression models 
controlled for differences in positions (15 C-level and VP -level position dummy variables), industry, geography, company maturity (both 
the number of rounds of financing raised and the number of employees), executive backgrounds (founder status, years of prior work 
experience, tenure in company, and educational background), and executive's equity stake. 

23. (Jensen etal, 1976) 

24. Using a sample of executive compensation data for 316 information technology firms, Anderson et al (2000) found that 
employers substitute bonuses for options and vice versa. 



CHAPTER NEVE 



1. (Gorman etaL, 1989:241) 

2. Gimeno et aL (1997) used a sample from the National Federation for Independent Businesses database that included more than 
1,500 entrepreneurs whose businesses had been in operation for less than 18 months. 

3. For a discussion of how contextual factors and growth goals affect financing choices in startups, see Sapienza et al. (2003). 

4. More general^, agency problems — problems that arise from the separation of owners (the principals) and management (their 
agents), where the agents make self-interested decisions rather than decisions that the owners would make themselves — can be reduced 
by the use of tangible assets that can be used as collateral, opening up new financing options (Jensen et al., 1976). 

5. In the Silicon Valley startups included in Stanford's SPEC project, the average startup required about $2.5 million in funding 
(Burton, 1995). 

6. Analyses of the first round of SPEC companies, which included 100 high-tech firms in Silicon Valley, found that firms having 
received VC financing were 47.7% more likely to undertake a series of human resources management (HRM) policies that rationalized 
employment practices (Hellman et al, 2002). 

7. In their book The Venture Capital Cycle, Paul Gompers and Josh Lerner classify seed/startup rounds as "early rounds"; first and 
early-stage rounds as "middle rounds"; and second, third, expansion, and bridge rounds as "late rounds," with VCs increasingly focusing 
on late-stage investments (Gompers et al., 1999:183-187). Similar three-stage models have been proposed by others (e.g., Prowse, 1998; 
Sohl, 2003). 

8. For the A-round, these numbers vary slightly between the technology and life sciences industries, with 31% of founders investing 
in the A-round in technology startups and 36% in life sciences startups. For all subsequent rounds, the two industries are within two 
percentage points of each other. 

9. For more details, see Aldrich et al (1996). In contrast, regarding the other investors examined below, founders have a wide set of 
weaker ties that are of short duration, intermittent, and tapped when needed or for instrumental reasons. 

10. See Nanda (2008) for a rigorous examination of how capital availability can lead founders to start marginal businesses that 
wouldn't be founded when capital is more constrained. He concludes that wealthy individuals can start businesses with lower growth 
potential because they do not face "the discipline of external finance." 

11. httpy/www.hbs.edu/entrepreneurs/scottcoolchtml, accessed October 2010. 

12. (Heskett, 1996:2) 

13. (Cespedes, 2009:1) 

14. Press release. University of New Hampshire, Center for Venture Research, June 1 1, 2003. 

15. For more details, see Kerr et al (2010). 

16. (Prowse, 1998) 

17. For more details, see Sohl (2001-2009). 

18. (Kerr et al, 2010) Based on quantitative analyses of the investments made by the two angel forums. 

19. For more details, see Sohl (2001-2009). 

20. VC data are taken from Thomson Reuters's ThomsonOne Banker database, accessed September 20, 2010, for all venture 
capital deals in all industries. 

21. For more details, see Sohl (2001-2009). 

22. For more details, see Hall et al. (2010). In comparison, a 1995 study of 794 startups that had received VC financing between 
1961 and 1992 found that 23.8% of the startups had merged or been acquired, 22.5% had gone public, 15.6% had been liquidated or had 
gone bankrupt, and 38.1% remained private (Gompers, 1995). 

23. This study examined 246 offers to 149 startups in MIT's Entrepreneurship Laboratory (E-Lab). For more details, see Hsu 
(2004). 

24. (Gompers etal, 1999) 

25. The sample was taken from the VentureOne database from 1987 to 2000, including 16,613 financing rounds for 7,765 companies 
(Cochrane, 2004). 

26. VC data are taken from Thomson Reuters's ThomsonOne Banker database, accessed September 20, 2010, for all venture 
capital deals in all industries. 

27. (Stinchcombe, 1965) 

28. In a sample of 62 VCs that self-reported the activities they undertook for their portfolio investments, equity financing averaged a 
score of 3.63 out of 5, and interfacing with other investors averaged 3.62 out of 5 (MacMillan et al., 1989). 

29. Rosenstein et al.'s (1993) survey of CEOs of 162 high-tech companies included a list of 11 activity areas that was based on the 
survey used by MacMillan et al (1989). 

30. Corporate — or "strategic" — investors are usually larger corporations in the same or a related sector that invest in startups for 
strategic and/or financial reasons and may later become leading candidates to acquire the startup. 

31. "Prominence" was measured as the proportion of all biotech alliances in the industry in which each investor had participated. For 
more details, see Stuart et al. (1999). 



32. On a 5-point Likert-type scale, VCs reported they were most directly involved in "serving as a sounding board" to the top-team, 
with an average score of 3.77 out of 5 (MacMiUan et al., 1989). 

33. In a sample of 162 firms with VC financing, 38.3% of the sample rated serving as a sounding board as one of the top-three most 
helpful activities by their investors (Rosenstein et aL, 1993). 

34. (Hellman et al., 2002) 

35. In a study of 246 offers to 149 startups in MIT's E-Lab, the average entrepreneur with multiple offers gave up $4. 1 million in 
pre-money valuation in order to get a VC who would add value (Hsu, 2004). 

36. (Bussgang, 2010:114) 

37. As a proxy for the amount of time the board members informally spent on the startup's business, we used the number of hours 
per months that the board's nonexecutive chairman — the most active outside director — worked on the startup. 

38. These findings parallel findings from the literature on mentor-protege relationships, which suggest that inexperienced executives 
receive less mentorship than they need because they have lower readiness for mentorship (Healy et al., 1990), middle-experience 
executives get more mentorship (closer to the amount they need), and those with a lot of experience get less mentorship because they 
require less (Kram, 1985). 

39. Chief executives with backgrounds in sales and other output functions may have more ability to grow the business beyond its 
earliest stages than do CEOs coming from throughput functions (Boeker et al., 2005). 

40. (Bagleyetal., 2003:112) 

41. The sample included 307 private biotechnology startups that received VC financing between 1978 and 1989 (Lemer, 1995). 

42. (Kaplan et al., 2004) 

43. (Wilmerding, 2004) 

44. This term actually has two parts. The "preference" governs whether investors recoup 100% of their investment, or a multiple 
thereof, before nonpreferred shareholders (e.g., the founders who own common stock) receive any proceeds. The second part is 
whether the new shares issued to the new investors are "participating" or not. Participation indicates whether or not investors participate 
in the back end of payouts, when the proceeds from the exit are divided among the common shareholders. The participation can also be 
capped at a certain level For more details, see Wasserman et al. (2010b). In a representative sample of VC financing deals that was 
compiled by the law firm Wilson Sonsini Goodrich & Rosati, 21%-29% of the investments included capped participation, 30%-35% 
included uncapped participation, and 40%-44% did not include any back-end participation for VCs. For details, see Baudler (2009). 

45. The more external risks the startup faces (e.g., the more uncertain its market size, expected customer adoption rate, competition, 
and financial market/exit condition risk), the more likely the VC will insist on terms that affect exit payoffs, such as redemption rights 
(Kaplan et al., 2004; Wilmerding, 2004). 

46. The fact that taking investor money increases organizational survival while increasing team turnover might suggest that the 
turnover that occurs may be beneficial by helping increase survival, a testable proposition that deserves systematic attention from 
researchers. 

47. Another 8% of seats were held by nonfounding executives, who may at times side with the founders and at other times side with 
the investors who helped hire them into the startup. 

48. In their analyses of the first round of SPEC firms, the increased use of HRM techniques in VC-backed startups led Hellman et 
al. (2002) to conclude that VCs play a critical role in "professionalizing" startups. 

49. VCs themselves often use the following line about the perils of having too many VCs on a startup's board: "VCs on your board 
are like martinis: One is good, two is better, three is a problem." 

50. Interestingly, one of the few potential advantages of receiving a down round may be the chance to shrink the startup's board. 
Venture capitalist Scott Yaphe observed, "When a company has raised capital through multiple rounds, and then undergoes a 
recapitalization or washout round, the board size tends to shrink. Investors often use the recap as a cleansing event and included in this 
cleansing is board composition. Non-participating or weak investor groups may be taken off the board and not replaced. The larger 
investors view a smaller board as a way to more efficiently control a bad situation that they are trying to improve." 

51. In a random sample of 794 VC-backed firms from 1961 to 1992, Gompers (1995) found that VCs used smaller/shorter rounds 
for startups where there was more uncertainty. 

52. Additional details about these secondary terms and how they affect both ownership and control can be found in Wasserman et 
al. (2010b). 

53. Standard investment agreements assume that founders are agents whose actions have to be closely monitored and who need 
strong incentives to act in the interest of the startup. However, to the extent that founders may sometime act more like stewards 
(Donaldson et al., 1991) — for whom such incentive structures may be unnecessary or even counterproductive, not to mention costly — 
their investment agreements might need to have different incentive structures than those used with agents. 



CHAPTER TEN 



1. In 2008 and 2009, 1 enriched my annual survey to ask more detailed questions about founder-CEO succession. These surveys are 
the source of the data in this chapter regarding (a) triggers, (b) whether the replaced founder stayed on as an executive or a board 
member, and (c) the background of the successor-CEOs. These data include 169 founder-CEO succession events. 

2. (Viranyetal., 1986) 

3. (Gompers et al., 1999; Stuart et aL, 1999) 

4. (Wasserman, 2003) 

5. (Greiner, 1972; Kazanjian, 1988) 

6. (Emerson, 1962; Pfeffer et al., 1978) 

7. (Boeker et aL, 2002) 

8. (Agrawaletal.,2006) 

9. This is true regarding stock returns (Fee et aL, 2003) and operating performance (Agrawal et aL, 2006); bad performance in either 
metric dramatically increases the chances that the board will hire an outside successor. 

10. In analyses of the SPEC sample of high-tech startups. Burton et aL (2007) found that position successors (ie., the second 
generation or more of employees at companies) who followed "atypical" position creators (defined as individuals with functional 
backgrounds that differ from the average functional background for their position in the sample) had 15% to 23% higher turnover rates. 

11. (Hamermesh, 2003:7) 

12. (Hamermesh, 2003:7) 

13. (Hamermesh, 2003:8) 

14. (Kirsner, 2010) 



CHAPTER ELEVEN 



1. (Stevenson etaL, 1990:23) 

2. Evans et al. (1989:824) state that across their full population of entrepreneurs and non-entrepreneurs, "The average probability of 
being a constrained entrepreneur is 3.75 percent, and the average probability of being an unconstrained entrepreneur is 0.06 percent." 

3. (Starr et al, 1990; Stevenson et al, 1990; Venkataraman, 1997) 

4. (Romanelli, 1989:375) 

5. (Stinchcombe, 1965) 

6. (Peteraf, 1993) 

7. (Pfeffer et al, 1978) 

8. (e.g., Amit et al., 1990; Coff, 1999) 

9. (Hellman, 1998; Wasserman, 2003) 

10. (Lemer, 1995; Sahlman, 1990) 

11. For more details, see Wasserman (2006a). Past studies often used ownership percentage as a proxy for decision-making control 
(e.g., Finkelstein et al., 1989; Gomez-Mejia et al., 1987; Salancik et al, 1980). However, the startup context is one in which ownership 
and decision-making control can diverge markedly (Kaplan et al, 2003). Therefore, in my analyses, control of decision making at the 
CEO and board levels was observed directly rather than inferred from ownership percentages. 

12. (Miller, 2010) 

13. Erick Schonfeld, http://techcrunch.com/201 1/03/03/jack-dorsey-twitter-punched-stomach/, March 3, 201 1. 

14. (Miner, 2010) 

15. (Bhide,2000) 

16. (Wasserman, 2003) 

17. Brian Bell, chairman of Buffalo Angel Network, comments made March 1, 2006. 

18. Or, "Queen-oriented." 

19. (Baum et al, 2001) Likewise, large companies that do not pick a clear strategy — e.g., being best at cost leadership, product 
differentiation, or focus — are usually at a competitive disadvantage (Porter, 1980). 

20. (Marlin et al, 1994; Porter, 1980) 

21. (Hamermesh, 2003:9) 

22. (Abetti2005) 

23. (Hamermesh, 2003:4) 

24. (Bhide, 2000:149) 

25. For more details, see Sapienza et al (2003). 

26. However, three characteristics of a transaction — its frequency, uncertainty, and asset specificity (measured by the forgone 
economic benefits of discontinuing a relationship) — are positively related to the likelihood that a firm will bring the needed resource inside 
its own boundaries. This is because the hierarchical relationships within firms are more effective than market relationships at resolving 
potential disputes (Williamson, 1985). Further insights come from the "incomplete contracts" approach to deciding which assets belong 
within the firm's boundaries (e.g., Grossman et al, 1986; Hart et al., 1990). 

27. (Agrawal et al., 1995; Lillis et al., 1976) 

28. (Burt, 1997; Lin, 1999:35) 

29. (Burt, 1997) 

30. (Shane et al, 2002) 

31. (Burt, 1997) 

32. (Shane et al, 2002) 

33. In a detailed study of the New York City garment-manufacturing industry, Uzzi (1997) concluded that mutual reciprocity and 
dense networks allowed managers to enter into partnerships with suppliers without their having to waste time on additional due diligence. 

34. (Fredricksonetal, 1988) 

35. (Tsaietal., 1998) 

36. (Lippman et al., 2003) 

37. (Evans et al, 1989) 

38. (Holtz-Eakin et al, 1994) 

39. (e.g., Spence, 2002) 

40. In a sample of 8,753 startups from 1987 to 2000, Gompers et al. (2010) found that previous startup experience was associated 
with a 3.8% increase in successful IPO, and that previous experience leading a startup to IPO was associated with a 7.8% increase in 
likelihood of taking the current startup to IPO. 

41. (e.g., Finkelstein, 1992) 

42. (Mills, 1956:162) 

43. (Jensen etal, 1976) 



44. (e.g., Davis etaL, 1997) 

45. (e.g., Hamilton, 2000; Hurst et al, 2010b; Moskowitz et aL, 2002) 

46. (Stevenson et al, 1990) 

47. Ostrander (1987) examined established, endowed, private social-service agencies that serve families in a large U.S. city and 
depend on private funds for discretionary income. The actions of the agency leaders were constrained by the elite board members who 
provided funding to the agency. 

48. For more details on the GEM efforts and the reports they produce, see httpy/www3.babson.edu/eship/research- 
publications/gem. cfm. 

49. The latter two exceptions were France, with a wealth/control breakdown of about 10%/90%, and Switzerland, with a breakdown 
of about 22%/78%. 

50. For instance, see Gompers et al (2010). 

51. One example of such a partnership may be the franchising relationship, in which the founder of the franchisor decides not to 
grow by opening new company-owned stores (which often require a lot of capital) but instead to partner with franchisees who will 
provide the capital and personnel to run the new stores and who will give the franchisor a share (but certainly not all) of the financial 
gains. Using a "Rich vs. King" lens to examine franchising relationships may provide further insights into potential "75% Rich, 75% 
King" relationships with outside resource providers. 

52. (Aldrich et al., 2006) 

53. (Lemer, 1995) 

54. See Broughman et al (2010) on founders' exit issues. 

55. (Livingston, 2007:123) 

56. (Miller, 2010) 

57. (Black etal., 1995) 

58. (Chemmanur etal, 1999) 

59. (Black etal., 1995) 

60. (Lin et al., 1995) 

61. (Delfassy et al, 2010:7) 



APPENDIX A 

1. (Jensen etal., 1990) 

2. (Beattyetal, 1994) 

3. (Deckop, 1988; Henderson et al., 1996) 

4. (Wasserman, 2003, 2006a, 2006b; Wasserman et al., 2005; Wasserman et al., 2008; Wasserman et al., 2010a) 

5. (Wasserman, 2003, 2006b) 

6. (Wasserman, 2006b) 

7. (Finkelstein, 1992; Waldman et al., 2001) 



BIBLIOGRAPHY 



AbettiPA. 2005. The creative evolution of Steria. Creativity and Innovation Management 14(2): 191-204. 

Adams JS. 1965. Inequity in social exchange. In L Berkowitz (Ed.), Advances in Experimental Social Psychology, Vol. 2: 267-299. 
Academic Press: New York. 

AgrawalA, Knoeber CR, Tsoulouhas T. 2006. Are outsiders handicapped in CEO successions? Journal of Corporate Finance 12(3): 
619-644. 

Agrawal D, Lai R. 1995. Contractual arrangements in franchising: An empirical investigation. Journal of Marketing Research 
22(May): 213-221. 

Alchian AA, Demsetz H. 1972. Production, information costs and economic organization. American Economic Review 62: 777-795. 
Aldrich H. 2010. Beam me up, Scott(ie)! Institutional theorists' struggles with the emergent nature of entrepreneurship. In W Sine, R 

David (EdiS.), Research in the Sociology of Work, on Institutions and Entrepreneurship , Vol. 21: 329-364. Emerald: Bradford, 

UK. 

Aldrich H, Carter N, Ruef M. 2004. Teams. In WB Gartner, KG Shaver, P Reynolds (Eds.), Handbook of Entrepreneurial 

Dynamics: 299-324. Sage: Thousand Oaks, CA. 
Aldrich H, Ruef M. 2006. Organizations Evolving (2nd ed.). Sage: Thousand Oaks, CA. 

Aldrich HE, Elam AB, Reese PR. 1996. Strong ties, weak ties, and strangers. In S Birley, IC MacMiilan (Eds.), Entrepreneurship in a 

Global Context: 1-25. Routledge: London. 
Allen P. 2011. Idea Man. Portfolio/Penguin, New York. 

Amason AC. 1996. Distinguishing the effects of functional and dysfunctional conflict on strategic decision making: Resolving a paradox 

for top management teams. Academy of Management Journal 39: 123-149. 
Amason AC, Sapienza HJ. 1997. The effects of top management team size and interaction norms on cognitive and affective conflict. 

Journal of Management 23(4): 495-516. 
Amit R, Glosten L, MuUer E. 1990. Entrepreneurial ability, venture investments, and risk sharing. Management Science 36(10): 1232- 

1245. 

Amit R, MacCrimmon KR, Zietsma C, Oesch JM. 2000. Does money matter? Wealth attainment as the motive for initiating growth- 
oriented technology ventures. Journal of Business Venturing 16: 119-143. 

Anderson MC, Banker RD, Ravindran S. 2000. Executive compensation in the IT industry. Management Science 3: 530-547. 

Austin W. 1980. Friendship and fairness: Effects of type of relationship and task performance on choice of distribution rules. Personality 
and Social Psychology Bulletin 6: 402-408. 

Awad J, Perkins B, Neff M. 2010. Building a team and Rich vs. King in the context of functional beverage ventures. Founders' 
Dilemmas Paper (April). 

Bagley CE, Dauchy CE. 2003. The Entrepreneur's Guide to Business Law. Thomson Learning: Ontario, Canada. 
Baker T, Aldrich H. 1994. Friends and strangers: Early hiring practices and idiosyncratic jobs. Frontiers of Entrepreneurship Research 
13: 75-87. 

Baker T, Miner AS, Lesley DT. 2003. Improvising firms: Bricolage, account-giving, and improvisational competencies in the founding 

process. Research Policy 32: 255-276. 
Baker T, Nelson RE. 2005. Creating something from nothing: Resource construction through entrepreneurial bricolage. Administrative 

Science Quarterly 50: 329-366. 

Baron J, Bielby T. 1986. Men and women at work: Sex segregation and statistical discrimination. American Journal of Sociology 91(4): 
759-799. 

Baron JN, Burton MD, Hannan MT. 1996. The road taken: Origins and evolution of employment systems in emerging companies. 

Industrial and Corporate Change 5(2): 239-275. 
Baron JN, Hannan MT. 2002. Organizational blueprints for success in high-technology start-ups: Lessons from SPEC. California 

Management Review 44(3): 8-36. 
Baron JN, Haiman MT, Burton MD. 2001. Labor pains: Change in organizational models and employee turnover in young, high-tech 

firms. American Journal of Sociology 106(4): 960-1012. 
Baron RA. 1998. Cognitive mechanisms in entrepreneurship: Why and when entrepreneurs think differently than other people. Journal 

of Business Venturing 13: 275-294. 
Baron RA, Ensley MD. 2006. Opportunity recognition as the detection of meaningful patterns: Evidence from novice and experienced 

entrepreneurs. Management Science 52(6): 1331-1344. 
Baudler M. 2009. From the WSGR Database: Financing trends. In The Entrepreneur's Report: Private Company Financing Trends . 

httpy/www.wsgr.com/publications/PDFSearch/entreport/Summer2009/private-company-financing-trends.htm#3. 



Baum JR, Bird BJ. 2010. The successful intelligence of high-growth entrepreneurs: Links to new venture growth. Organization Science 
21(2): 397-412. 

Baum JR, Locke EA. 2004. The relationship of entrepreneurial traits, skiU, and motivation to subsequent venture growth. Journal oj 

Applied Psychology 89(4): 587-598. 
Baum JR, Locke EA, Smith KG. 2001. A multidimensional model of venture growth. Academy of Management Journal 44(2): 292- 

303. 

Beatty RP, Zajac EJ. 1994. Managerial incentives, monitoring, and risk bearing: A study of executive compensation, ownership, and 

board structure in initial public offerings. Administrative Science Quarterly 39(2): 313-335. 
Beckes-GeUner U, Moog PM. 2008. Who chooses to become an entrepreneur? The jacks-of-all-trades in human and social capital 

University of Zurich Working Paper. 
Beckman C. 2006. The influence of founding team company affiliations on firm behavior. Academy of Management Journal 49(4): 

741-758. 

Bernardo AE, Welch 1. 2001. On the evolution of overconfidence and entrepreneurs. Yale International Center for Finance, Working 
Paper Series No. 137. 

Bhattacharyya S, Lafontaine F. 1995. Double-sided moral hazard and the nature of share contracts. RAND Journal of Economics 
26(4): 761-781. 

Bhide A. 1994. How entrepreneurs craft strategies that work. Harvard Business Review 72(2): 150-161. 
Bhide A. 2000. The Origin and Evolution of New Businesses. Oxford University Press: Oxford. 

Black BS, Gilson RJ. 1995. Venture capital and the structure of capital markets. Journal of Financial Economics 47(3): 243-277. 
Blanchfiower DG, Oswald AJ. 1998. What makes an entrepreneur? Journal of Labor Economics 1(1): 26-60. 
Blank S. 2009. Lies entrepreneurs tell themselves. httpy/steveblank.com/2009/06/15/lies-entrepreneurs-teU-themselves/. 
Blau P. 1970. A formal theory of differentiation in organizations. American Sociological Review 35: 201-218. 
Blau PM, Scott WR. 1962. Formal Organizations. Chandler: San Francisco. 

Boeker W. 1988. Organizational origins: Entrepreneurial and envirorunental imprinting at the time of founding. In GR CarroU (Ed.), 

Ecological Models of Organizations: 33-51. BaUinger: Cambridge, MA. 
Boeker W. 1989. Strategic change: The effects of founding and history. Academy of Management Journal 32: 489-515. 
Boeker W, Karichalil R. 2002. Entrepreneurial transitions: Factors influencing founder departure. Academy of Management Journal 

45(3): 818-826. 

Boeker W, WUtbank R. 2005. New venture evolution and managerial capabilities. Organization Science 16(2): 123-133. 
Broughman B, Fried J. 2010. Renegotiation of cash flow rights in the sale of VC-backed firms. Journal of Financial Economics 95(3): 
384-399. 

Bruderl J, Preisendorfer P, Ziegler R. 1992. Survival chances of newly founded business organizations. American Sociological Review 
57(April): 227-242. 

Burt RS. 1992. Structural Holes: The Social Structure of Competition. Harvard University Press: Cambridge, MA. 
Burt RS. 1997. The contingent value of social capital. Administrative Science Quarterly 42(2): 339-365. 

Burton D. 1995. The evolution of employment systems in high technology firms. Unpublished Ph.D. dissertation. Department of 
Sociology, Stanford University. 

Burton MD, Beckman C. 2007. Leaving a legacy: Position imprints and successor turnover in young firms. American Review of 
Sociology 72: 239-266. 

Busenitz LW, Barney JB. 1997. Differences between entrepreneurs and managers in large organizations: Biases and heuristics in 

strategic decision-making. Journal of Business Venturing 12(1): 9-30. 
Bussgang J. 2010. Mastering the VC Game. Portfolio hardcover. 

Carland JW, Hoy F, Boulton W, Carland JAC. 1984. Differentiating entrepreneurs from small business owners: A conceptualization. 

Academy of Management Review 9: 354—359. 
Carpenter M, Wade J. 2002. Microlevel opportunity structures as determinants of non-CEO executive pay. Academy of Management 

Journal 45(6): 1085-1103. 
Cespedes FV. 2009. Intuit. Harvard Business School Case No. 810-018. 

Chemmanur TJ, Fulghieri P. 1999. A theory of the going-public decision. Review of Financial Studies 12(2): 249-279. 
Christensen CM, Bower JL. 1996. Customer power, strategic investment, and the failure of leading firms. Strategic Management 
Journal 17: 197-218. 

Cochrane JH. 2004. The Risk and Return of Venture Capital. University of Chicago: Chicago. 

Coff RW. 1999. When competitive advantage doesn't lead to performance: The resource-based view and stakeholder bargaining power. 

Organization Science 10(2): 119-133. 
Coleman JS. 1988. Social capital in the creation of human capital American Journal of Sociology 94: 95-120. 

Cooper AC, Woo CY, Dunkelberg WC. 1988. Entrepreneurs' perceived chances for success. Journal of Business Venturing 3: 97- 
108. 

Davidsson P, Honig B. 2003. The role of social and human capital among nascent entrepreneurs. Journal of Business Venturing 18: 
301-331. 



Davis JH, Schoorman FD, Donaldson L. 1997. Toward a stewardship theory of management. Academy of Management Review 22(1): 
20-47. 

Deckop JR. 1988. Determinants of chief executive officer compensation. Industrial and Labor Relations Review 41(2): 215-226. 
Delfassy N, Price G. 2010. Entrepreneurial contradictions: Making sense of differences in theory and practice. Founders' Dilemmas 
Paper (April). 

Deutsch M. 1975. Equity, equality and need: What determines which value will be used as the basis for distril^utive justice? Journal oj 
Social Issues 31: 137-149. 

Dobrev SD, Barnett WP. 2005. Organizational roles and transition to entrepreneurship. Academy of Management Journal 48(3): 433- 
449. 

Donaldson L, Davis JH. 1991. Stewardship theory or agency theory: CEO governance and shareholder returns. Australian Journal of 
Management 16(1): 49-64. 

Economides N. 1996. The economics of networks. International Journal of Industrial Organization 14(6): 673-699. 
Edmondson AC, McManus SE. 2007. Methodological fit in management field research. Academy of Management Review 32(4): 1155- 
1179. 

Eisenhardt KM, Bourgeois III L. 1988. Politics of strategic decision making in high-velocity environments. Academy of Management 
Journal 31: 143-159. 

Eisenhardt KM. 1989. Making fast strategic decisions in high-velocity environments. Academy of Management Journal 32(3): 543- 
576. 

Eisenhardt KM, Schoonhoven CB. 1990. Organizational growth: Linking founding team, strategy, environment, and growth among U.S. 

semiconductor ventures, 1978-1988. Administrative Science Quarterly 35: 504—529. 
Eisenmann TR, Parker G, Alstjme MV. 2006. Strategies for two-sided markets. Harvard Business Review 84(10): 92-101. 
EUenbein DW, Hamilton BH, Zenger TR. 2009. Entrepreneurial spawning of scientists and engineers: Stars, slugs and small firms. 

Washington University, St Louis Working Paper. 
Emerson RM. 1962. Power-dependence relations. American Sociological Review 27: 31-40. 

Evans DS, Jovanovic B. 1989. An estimated model of entrepreneurial choice under liquidity constraints. Journal of Political Economy 
97(4): 808-827. 

Fama EF, Jensen ML. 1983. Separation of ownership and control. Journal of Law and Economics 26: 301-325. 
Fee CE, Hadlock C. 2003. Raids, rewards, and reputations in the market for managerial talent. Review of Financial Studies 16(4): 
1315-1357. 

Fernandez R, Sosa ML. 2005. Gendering the job: Networks and recruitment at a call center American Journal of Sociology 111(3): 
859-904. 

Finkelstein S. 1992. Power in top management teams: Dimensions, measurement, and validation. Academy of Management Journal 35: 
505-538. 

Finkelstein S, Hambrick DC. 1989. Chief executive compensation: A study of the intersection of markets and political processes. 

Strategic Management Journal 10: 121-134. 
Fiske ST, Taylor SE. 1991. Social Cognition: From Brains to Culture. McGraw-Hill: New York. 

Florin J, Lubatkin M, Schuke W. 2003. A social capital model of high-growth ventures. Academy of Management Journal 46(3): 374— 
384. 

Fredrickson JW, Hambrick DC, Baumrin S. 1988. A model of CEO dismissal Academy of Management Review 13(2): 255-270. 
Galbraith J. 1982. The stages of growth. Journal of Business Strategy (Summer): 70-79. 

Gimeno J, Folta TB, Cooper AC, Woo CY. 1997. Survival of the fittest? Entrepreneurial human capital and the persistence of 

underperforming firms. Administrative Science Quarterly 42(4): 750-783. 
Goldin C. 2008. Gender gap. In Library of Economics and Liberty: The Concise Encyclopedia of Economics. 

httpy/www.econHb.org/library/Enc/GenderGap.htmL 
Gomez-Mejia LR, Tosi HL, Hinken T. 1987. Managerial control, performance, and executive compensation. Academy of Management 

Journal 30:51-70. 

Gompers P. 1995. Optimal investment, monitoring, and the staging of venture capital Journal of Finance 50: 1461-1489. 
Gompers P, Lerner J. 1999. The Venture Capital Cycle. MIT Press: Boston. 

Gompers P, Lemer J, Scharfstein D. 2005. Entrepreneurial spawning: Public corporations and the genesis of new ventures. Journal of 
Finance 60(2): 577-614. 

Gompers P, Lemer J, Scharfstein D, Kovner AR. 2010. Performance persistence in entrepreneurship and venture capital. Journal of 

Financial Economics 96(1): 18-32. 
Gorman M, Sahlman WA. 1989. What do venture capitalists do? Journal of Business Venturing 4(4): 231-248. 
Granovetter M. 1973. The strength of weak ties. American Journal of Sociology 78:1360-1380. 
Greiner LE. 1972. Evolution and revolution as organizations grow. Harvard Business Review 50(4): 37-46. 

Grossman SJ, Hart OD. 1986. The costs and benefits of ownership: A theory of vertical and lateral integration. Journal of Political 
Economy 94(4): 691-719. 

Grossman SJ, Hart OD. 1988. One share-one vote and the market for corporate control. Journal of Financial Economics 20: 175-202. 



Groysberg B. 2010. Chasing Stars: The Myth of Talent and the Portability of Performance. Princeton University Press: Princeton, 
NJ. 

Hackman JR, Wageman R, Ruddy T, Ray CR. 2000. Team effectiveness in theory and practice. In E Locke (Ed.), Industrial and 

Organizational Psychology: Linking Theory with Practice: 275-294. Blackwell: London. 
Haleblian J, Finkelstein S. 1993. Top management team size, CEO dominance, and i3rm performance: The moderating roles of 

environmental turbulence and discretion. Academy of Management Journal 38(4): 844-863. 
Hall RE, Woodward SE. 2010. The burden of the nondiversifiable risk of entrepreneurship. American Economic Review lOO(June): 

1163-1194. 

Hambrick DC, Mason PA. 1984. Upper echelons: The organization as a reflection of its top managers. Academy of Management 
Review 9:193-206. 

Hamermesh RG. 2003. Managing Segway's early development. Harvard Business School Case No. 804-065. 
Hamermesh RG, Barley L, Graham G. 2008. Proteus Biomedical: Making pigs fly. Harvard Business School Case No. 809-051. 
Hamilton BH. 2000. Does entrepreneurship pay? An empirical anafysis of the returns to self-employment. Journal of Political 
Economy 108(3): 604-631. 

Hannan MT, Burton D, Baron J. 1996. Inertia and change in the early years: Employment relations in young, high-tech firms. Industrial 

and Corporate Change 5(2): 239-275. 
Hannan MT, Freeman J. 1989. Organizational Ecology. Harvard University Press: Cambridge, MA. 

Hart MM, Roberts MJ, Stevens JD. 2003. Zipcar: Refining the business model Harvard Business School Case No. 803-096. 
Hart O, Moore J. 1990. Property rights and the nature of the firm. Journal of Political Economy 98(6): 1119-1158. 
Hayward M, Shepherd DA, Griffin D. 2006. A hubris theory of entrepreneurship. Management Science 52(2): 160-172. 
Healy CC, Welchart A. 1990. Mentoring relations: A definition to advance research and practice. Educational Researcher 19(9): 17- 
21. 

Hegedus C, Noujaim J. 2001. Startup.com (Documentary). Lrve/Artisan: Santa Monica, CA. 

Hellman T. 1998. The allocation of control rights in venture capital contracts. RAND Journal of Economics 29: 57-76. 
Helhnan T, Puri M. 2002. Venture capital and the professionalization of start-up firms: Empirical evidence. Journal of Finance 57(1): 
169-197. 

Henderson AD, Fredrickson JW. 1996. Information-processing demands as a determinant of CEO compensation. Academy of 

Management Journal 39(3): 575-606. 
Heskett JL. 1996. Scott Cook and Intuit. Harvard Business School Case No. 396-282. 

Higgins M. 2005a. Building career foundations: Humphrey Chen. Harvard Business School Video Series No. 405-704. 

Higgins M. 2005b. Career Imprints: Creating Leaders across an Industry. Jossey-Bass: San Francisco. 

Higgins M, Wasserman N. 2010. Humphrey and Cecilia. Harvard Business School Video Series No. 810-702. 

Hills G, Singh R. 2004. Opportunity recognition. In WB Gartner, KG Shaver, P Reynolds (Eds.), Handbook of Entrepreneurial 

Dynamics: 259-272. Sage: Thousand Oaks, CA. 
Hitt MA, Bierman L, Shimizu K, Kochhar R. 2001. Direct and moderating effects of human capital on strategy and performance in 

professional service firms: A resource-based perspective. Academy of Management Journal 44(1): 13-28. 
Hmieleski KM, Baron RA. 2009. Entrepreneurs' optimism and new venture performance: A social cognitive perspective. Academy of 

Management Journal 52(3): 473-488. 
Holmstrom B. 1982. Moral hazard in teams. Bell Journal of Economics 13: 324—340. 

Holtz-Eakin D, Joulfaian D, Rosen HS. 1994. Entrepreneurial decisions and liquidity constraints. RAND Journal of Economics 25(2): 
334-347. 

Hsu D. 2004. What do entrepreneurs pay for venture capital affiliation? Journal of Finance 59(4): 1805-1844. 
Hsu D. 2007. Experienced entrepreneurs, capital and VC funding. Research Policy 36: 722-741. 

Hsu D, Roberts DR, Lesley D. 2007. Entrepreneurs from technology-based universities: Evidence from MIT. Organization Science 
6(3): 768-788. 

Hurst E, Li G, Pugsley B. 2010a. Using expenditures to estimate missing self employed income. University of Chicago Working 
Paper. 

Hurst EG, Pugsley BW. 2010b. Non pecuniary benefits of small business ownership. University of Chicago Working Paper. 

Jehn KA. 1997. Affective and cognitive conflict in work groups: Increasing performance through value-based intragroup conflict. In 

CKW De Dreu, E Van de Vliert (Eds.), Using Conflict in Organizations: 87-100. Sage: Thousand Oaks, CA. 
Jensen M, Meclding W. 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial 

Economics 3: 305-360. 

Jensen M, Murphy K. 1990. Performance pay and top management incentives. Journal of Political Economy 98: 225-264. 
Kabanoff B. 1991. Equity, equality, power and conflict. Academy of Management Review 16: 416-441. 

Kaplan SN, Stromberg P. 2003. Financial contracting theory meets the real world: An empirical anafysis of venture capital contracts. 

Review of Economic Studies 70: 281-315. 
Kaplan SN, Stromberg P. 2004. Characteristics, contracts, and actions: Evidence from venture capitalist anaty^ses. Journal of Finance 

59: 2173-2206. 



Kawasaki G. 2006. Top ten lies of entrepreneurs, http://blog.guykawasaki.com/2006/01/the_top_ten_lie_l.html. 

Kazanjian RK. 1988. Relation of dominant problems to stages of growth in technology-based new ventures. Academy of Management 
Journan\(2): 251-219. 

Keck SL. 1997. Top management team structure: Differential effects by environmental context. Organization Science 8(2): 143-156. 
Kerr W, Lerner J, Schoar A. 2010. The consequences of entrepreneurial finance: A regression discontinuity analysis. NBER Working 

Paper. Harvard Business School, MIT. 
Kim PH, Aldrich HE. 2005. Social capital and entrepreneurship. In Foundations and Trends in Entrepreneurship : Vol 1, Issue 2. 

NOW: Hanover, MA. 

Kirsner S. 2010. Legendary Hub venture investor, musician works on his next hit. Boston Globe. 

httpy/www.boston. com/bus iness/articles/2010/07/ 1 8/legendary_hub 

_venture_investor_musician_works_on_his_next_hit/. 
Kirzner IM. 1973. Competition and Entrepreneurship. University of Chicago Press: Chicago. 

Knight D, Pearce C, Smith KG, Olian JD, Sims HP, Smith KA, Flood P. 1999. Top management team diversity, group process and 

strategic consensus. Strategic Management Journal 20: 445-465. 
Kram KE. 1985. Mentoring at Work: Developing Relationships in Organizational Life. Scott, Foresman: Glenview, IL. 
Lawler EEL 1971. Pay and Organizational Effectiveness: A Psychological View. McGraw-Hill: New York. 
Lazear EP. 2004. Balanced skills and entrepreneurship. American Economic Review 94: 208-211. 

Lease RC, McConneU JJ, Mikkelson WH. 1983. The market value of control in publicly-traded corporations. Journal of Financial 

Economics 11:439-471. 
Lencioni P. 2002. The Five Dysfunctions of a Team. Jossey-Bass: San Francisco. 
Lerner J. 1995. Venture capitalists and the oversight of private firms. Journal of Finance 50: 301-318. 

Leventhal GS. 1976. The distribution of rewards and resources in groups and organizations. In L Berkowitz, E Walster (Eds.), Equity 

Theory: Toward a General Theory of Social Interaction: 92-133. Academic Press: New York. 
Leventhal GS, Karuza J, Fry WR. 1980. Beyond fairness: A theory of allocation preferences. In G Mikula (Ed.), Justice and Social 

Interaction: Experimental and Theoretical Contributions from Psychological Research: 167-208. Springer-Verlag: New York. 
Lillis C, Chem N, Gilman J. 1976. Competitive advantage variation over the life cycle of a franchise. Journal of Marketing 40(4): 77- 

80. 

Lin N. 1999. Building a network theory of social capital Connections 22(1): 28-51. 

Lin N, Ensel W, Vaughn JC. 1981. Social resources and strength of ties: Structural factors in occupational status attainment. American 

Sociological Review 46(August): 393-405. 
Lin TH, Smith RL. 1995. Insider reputation and selling decisions: The unwinding of venture capital investments during equity IPOs. 

Claremont College Working Paper. 
Lippman SA, Rumelt RP. 2003. A bargaining perspective on resource advantage. Strategic Management Journal 24: 1069-1086. 
Livingston J. 2007. Founders at Work. Apress: New York. 

Lovallo D, Kahneman D. 2003. Delusions of success: How optimism undermines executives' decisions. Harvard Business Review 
81(7): 5^67. 

MacMillan IC, Kulow DM, Khoylian R. 1989. Venture capitalists' involvement in their investment: Extent and performance. Journal of 
Business Venturing 4(1): 27-47. 

Malach-Pines A, Sadeh A, Dvir A, Yafe-Yanai O. 2002. Entrepreneurs and managers: Similar yet different. International Journal of 

Organizational Analysis 10(2): 172-190. 
Malhotra D. 2009. When contracts destroy trust. Harvard Business Review 87(5): 25. 

Marlin D, Lamont BT, Hoffman JJ. 1994. Choice situation, strategy, and performance: A reexamination. Strategic Management 
Journal 15: 229-239. 

McCallMW. 1998. High Flyers: Developing the Next Generation of Leaders. Harvard Business School Press: Boston. 

McDougall P, Oviatt B. 1996. New venture internationalization, strategic change, and performance: A follow-up study. Journal of 

Business Venturing 11(1): 23-40. 
Meyer GD, Dean TJ. 1990. Upper echelons perspective on transformational leadership problems in high-technology firms. Journal of 

High Technology Management Research 1(2): 223-242. 
Miller CC. 2010. Why Twitter's C.E.O. demoted himself. New York Times, October 30, 2010. 
Mills CW. 1956. The Power Elite. Oxford University Press: New York. 

Miner JB, Raju NS. 2004. Risk propensity differences between managers and entrepreneurs and between low- and high-growth 

entrepreneurs: A reply in a more conservative vein. Journal of Applied Psychology 89(1): 3-13. 
Moore GA. 2002. Crossing the Chasm (3rd ed.). HarperCollins: New York. 

Mosakowski E. 1998. Entrepreneurial resources, organizational choices, and competitive outcomes. Organization Science 9(6): 625- 
643. 

Moskowitz TJ, Vissing-Jorgensen A. 2002. The returns to entrepreneurial investment: A private equity premium puzzle? American 

Economic Review 92(4): 745-778. 
Nanda R. 2008. Entrepreneurship and the discipline of external finance. Harvard Business School Working Paper No. 08-047. 



Northcraft GB, Neale MA. 1987. Experts, amateurs, and real estate: An anchoring-and-adjustment perspective on property pricing 

decisions. Organizational Behavior and Human Decision Processes 39(1): 84—97. 
O'Reilly CA, Caldwell D, Bamett WP. 1989. Work group demography, social integration, and turnover Administrative Science 

Quarterly 34: 21-37. 

Ostrander SA. 1987. Elite domination in private social agencies. In GW Domhoff, TR Dye (Eds.), Power Elites in Organizations: 85- 
102. Sage: Thousand Oaks, CA. 

Pennings JM. 1982. Organizational birth frequencies: An empirical investigation. Administrative Science Quarterly 29(1): 120-144. 
Pennings JM, WezelFC. 2010. Faraway, yet so close: Organizations in demographic flux. Organization Science 21(2): 451-468. 
Peteraf MA. 1993. The cornerstones of competitive advantage: A resource-based view. Strategic Management Journal 14(3): 179- 
191. 

Pfeffer J, Salancik GR. 1978. The External Control of Organizations: A Resource Dependence Perspective . Harper and Row: New 
York. 

Porter ME. 1980. Competitive Strategy. Free Press: New York. 

Prowse S. 1998. Angel investors and the market for angel investments. Journal of Banking and Finance 22(6-8): 785-792. 
Reynolds PD, Curtin RT. 2008. Business Creation in the United States. NOW: Hanover, MA. 

Roberts P, Khaire M. 2009. Getting known by the company you keep: Publicizing the qualifications and former associations of skilled 

employees. Industrial and Corporate Change 18(1): 77-106. 
Rogers P, Blenko M. 2006. Who has the D? How clear decision roles enhance organization performance. Harvard Business Review 

84(1): 52-61. 

RomaneUi E. 1989. Environments and strategies of organization startup: Effect on early survival. Administrative Science Quarterly 
34(3): 369-387. 

Rosenstein J, Bruno AV, Bygrave WD, Taylor NT. 1993. The CEO, venture capitalists, and the board. Journal of Business Venturing 
8(2): 99-113. 

Ruef M. 2009. Economic inequality among entrepreneurs. In LA Keister (Ed.), Economic Sociology of Work, Vol. 18: 57-71. Emerald: 
Bradford, UK. 

Ruef M. 2010. The Entrepreneurial Group: Social Identities, Relations, and Collective Action . Princeton University Press: 
Princeton, NJ. 

Ruef M, Aldrich HE, Carter N. 2003. The structure of founding teams: Homophily, strong ties, and isolation among U.S. entrepreneurs. 

American Sociological Review 68: 195-222. 
Sabherwal R, Hirschheim R, Goles T. 2001. The dynamics of alignment: Insights from a punctuated equilibrium model Organization 

Science 12(2): 179-197. 

Sahknan W. 1990. The structure and governance of venture-capital organizations. Journal of Financial Economics 27: 473-521. 
Salancik GR, Pfeffer J. 1980. Effects of ownership and performance on executive tenure in U.S. corporations. ^caJemy of 

Management Journal 23(4): 653-664. 
Sapienza HJ, Korsgaard MA, Forbes DP. 2003. The self-determination motive and entrepreneurs' choice of financing. In JA Katz, D 

Shepherd (Eds.), Cognitive Approaches to Entrepreneurship Research, Vol 6: 105-138. Elsevier: Burlington, MA. 
Sarasvathy S. 2008. Effectuation: Elements of Entrepreneurial Expertise. Edward Elgar: Northampton, MA. 

Schaubroeck J, Ganster DC, Jones JR. 1998. Organization and occupational influences in the attraction-selection-attribution process. 

Journal of Applied Psychology 83: 869-891. 
Schumpeter J. 1934. Theory of Economic Development. Harper and Row: New York. 
Schumpeter J. 1942. Capitalist, Socialism, and Democracy. Harper & Brothers: New York. 
Scitovszky TD. 1943. A note on profit maximization and its implications. Review of Economic Studies 1 1(1): 57-60. 
Shane S. 2008. Illusions of Entrepreneurship. Yale University Press: New Haven, CT. 

Shane S, Stuart T. 2002. Organizational endowments and the performance of university start-ups. Management Science 48(1): 154—170. 
Shane S, Venkataraman S. 2000. The promise of entrepreneurship as a field of research. Academy of Management Review 25: 217- 
226. 

Sherer PD. 1995. Leveraging human assets in law firms: Human capital structures and organizational capabilities. Industrial and Labor 

Relations Review 48: 671-691. 
Singh JV, Lumsden CJ. 1990. Theory and research in organizational ecology. Annual Review of Sociology 16: 161-195. 
Sohl J. 2001-2009. Annual report: The angel investor market. Center for Venture Research, University of New Hampshire. 
Sohl JE. 2003. The U.S. angel and venture capital market recent trends and developments. Journal of Private Equity 6(2): 7-17. 
Spence M. 2002. Signaling in retrospect and the informational structure of markets. American Economic Review 92(3): 434—459. 
Spolsky J. 2003. Fixing venture capital Joel on Software blog, June 3, 2003. httpy/www.joelonsoftware.com/articlesA^C.htmL 
Stam E, Thurik R, Van der Zwan P. 2008. Entrepreneurial exit in real and imagined markets. Tinbergen Institute Discussion Paper 

No. 08-031/3. 

Starr JA, MacMillan IC. 1990. Resource cooptation via social contracting: Resource acquisition strategies for new ventures. Strategic 
Management Journal 11: 79-92. 

Stevenson HH, Jariilo JC. 1990. A paradigm of entrepreneurship: Entrepreneurial management. Strategic Management Journal 11(1): 



17-27. 

Stewart WH, Roth PL. 2001. Risk propensity differences between entrepreneurs and managers: A meta-analytic review. Journal oj 

Applied Psychology 86(1): 145-153. 
Stinchcombe AL. 1965. Organizations and social structure. In JG March (Ed.), Handbook of Organizations: 153-193. Rand McNally: 

Chicago. 

Strasser JB, Becklund L. 1991. Swoosh: The Unauthorized Story of Nike and the Men Who Played There. Harcourt Brace 
Jovanovich: New York. 

Stuart T, Hoang H, Hybels R. 1999. Interorganizational endorsements and the performance of entrepreneurial ventures. Administrative 
Science Quarterly 44: 315-349. 

Taylor A, Greve HR. 2006. Superman or Fantastic Four? Knowledge combination and experience in innovative teams. Academy of 

Management Journal 49(4): 723-740. 
Tsai W, Ghoshal S. 1998. Social capital and value creation: The role of intrafirm networks. Academy of Management Journal 41(4): 

464-476. 

Ucbasaran D, Lockett A, Wright M, Westhead R 2003. Entrepreneurial founding teams: Factors associated with member entry and exit. 

Entrepreneurs hip Theory and Practice 28: 107-128. 
Uzzi B. 1997. Social structures and competition in interfirm networks: The paradox of embeddedness. Administrative Science 

Quarterly 42(1): 35-67. 

Venkataraman S. 1997. The distinctive domain of entrepreneurship research, hi Advances in Entrepreneurship , Firm Emergence, and 

Growth, Vol 3: 119-138. JAl: Greenwich, CT. 
Virany B, Tushman ML. 1986. Top management teams and corporate success in an emerging industry. Journal of Business Venturing 

1(3): 261-274. 

Virany B, Tushman ML, RomaneUi E. 1992. Executive succession and organization outcomes in turbulent environments: An organization 

learning approach. Organization Science 3(1): 72-91. 
Wadhwa V, Aggarwal R, Holly K, Salkever A. 2009. The anatomy of an entrepreneur: Family background and motivation. Kauffman 

Foundation Research Paper (July). 
Waldinger RD, Lichter Ml. 2003. How the Other Half Works: Immigration and the Social Organization of Labor . University of 

California Press: Berkeley. 

Waldman DA, Ramirez GG, House RJ, Puranam P. 2001. Does leadership matter? CEO leadership attributes and profitability under 
conditions of perceived environmental uncertainty. Academy of Management Journal 44(1): 134—143. 

Walsh JP. 1995. Managerial and organizational cognition: A trip down memory lane. Organization Science 6(3): 280-321. 

Walton RE, McKersie RB. 1965. A Behavioral Theory of Labor Negotiations: An Analysis of a Social Interaction System. 
McGraw-Hill: New York. 

WassermanN. 2002. The venture capitalist as entrepreneur: Characteristics and dynamics within VC firms. Unpublished Ph.D. thesis, 
Harvard University. 

WassermanN. 2003. Founder-CEO succession and the paradox of entrepreneurial success. Organization Science 14(2): 149-172. 
Wasserman N. 2005. Upside-down venture capitalists and the transition toward pyramidal firms. In L Keister (Ed.), Research in the 

Sociology of Work: Entrepreneurship: 151-208. JAl: Greenwich, CT. 
WassermanN. 2006a. Rich vs. King: The entrepreneur's dilemma. Best Paper Proceedings, Academy of Management. 
WassermanN. 2006b. Stewards, agents, and the founder discount: Executive compensation in new ventures. Academy of Management 

Journal 49: 960-976. 

Wasserman N. 2008a. Don't wait too long to become an entrepreneur. Harvard Business Review Blogs, September 9, 2008. 

httpy/blogs. hbr.org/cs/2008/09/dont_wait_too_long_to_become_a. html 
Wasserman N. 2008b. Revisiting the strategy-structure -performance paradigm: The case of venture capital Organization Science 

19(2): 241-259. 

Wasserman N, Barley L. 2009. A note on the legal and tax implications of founders' equity splits. Harvard Business School Note No. 
809-110, 1-15. 

Wasserman N, Boeker W. 2005. Mentoring and monitoring: The evolution of boards of directors in new ventures. Best Paper 

Proceedings, Babson Research Conference. 
Wasserman N, Hellman T. 2010a. The first deal: The division of founder equity in new ventures. Academy of Management Annual 

Meeting. 

Wasserman N, Marx M. 2008. Split decisions: How social and economic choices affect the stability of founding. Academy of 
Management Annual Meeting. 

Wasserman N, Nazeeri F, Anderson K. 2010b. A "Rich- vs. -King" approach to term sheet negotiations. Harvard Business School Note 
No. 810-119. 

Weber M. 1946. From Max Weber: Essays in Sociology. Oxford University Press: New York. 
Weber M. 1949/1997. The Methodology of the Social Sciences (EA Shils, HA Finch, Trans.). Free Press: New York. 
Wiersema MF, Bantel KA. 1992. Top management team demography and corporate strategic change. Academy of Management 
Journal ?>5{l):9\-m. 



Williamson OE. 1985. The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. Free Press: New York. 
Wilmerding A. 2004. Term Sheets & Valuations. Aspatore: Boston. 

Wozniak S. 2006. iWoz: Computer Geek to Cult Icon: How I Invented the Personal Computer, Co-Founded Apple, and Had Fun 
Doing It. Norton: New York. 

Young JS, Simon WL. 2004. iCon: Steve Jobs, the Greatest Second Act in the History of Business. John Wiley: San Francisco. 
Zaleznik A, Kets de Vries MFR. 1975. Power and the Corporate Mind. 
Houghton Mifflin: Boston. 



INDEX 



Abbott Labs, 45 

accountability and role assignments, 128 
acquisition route, exit decisions, 376-83 
Adams, Tim, 320, 346 

Addante, Frank: decision-making approach, 191 

entrepreneurial motivations, 33, 119, 355 

equity split decisions, 150, 151, 162, 197 

financing decisions, 252, 271-72 

hiring decisions, 228-29, 230, 343-44 

IPO decision-making, 383, 384, 385 
Adnet, 150 

age factor, entrepreneurship decision, 27-29, 34-35, 427nn3-5 

agency theory, 180-82, 362-63, 447n53 

Aldrich, H., 6n, 428nl3, 432n50, 435n3 

The Alexander Group (TAG), 49-50 

alignment, decision-making. See Three Rs equilibrium 

Allen, Paul, 166 

altruism motivation, entrepreneurship, 32, 33, 35 

Anderson, M. C, 443n24 

angel investors: overview, 255, 262-67 

and entrepreneurship timing, 66 

financial capital statistics, 20, 270, 427n31 

social capital benefits, 271. See also financing approaches; venture capitalists 
Antje, at Zipcar, 167 

Apple Computer. See Jobs, Steve and Wozniak, Steve 
autocracy blueprint, hiring decisions, 210n 
autocratic model, decision-making 
approach, 129-30, 134-38, 437nn6, 9 
autonomy motivation, 33 



Bagley, C. E., 277 

Baker, T., 435n2, 441n3, 442nl2 

Balsillie, James, 132n 

Baron, J, 443n22 

Baron, J. N., 436nl9 

Baum, J. R.,433n59 

Baxter company, 44—45 

Beckes-Gellner, U., 435nl 1 

Ben and Jerry's, 13 In 

Bergen, Chris, 384 

Bernardo, A. E.,433n58 

Bhide, Amar, 3, 425n3 

Biblical wisdom, 99, 307n 

Blanchflower, D. G., 48 

Blank, Steve, 65 

Blau, P., 437nl4 

Blogger. See Williams, Evan 

blueprint study, hiring decisions, 210-11, 441nnl-2 (ch 8) 
Bbmberg, Matt, 287 

board of directors: angel investor participation, 263-64 
and founder backgrounds, 45-46 



founder participation numbers, 138-40, 143-44, 285 
solo approach correlations, 340 

venture capitalist impact, 273-78, 284-93, 446nn48-50, 447n51. See also succession process, founder-CEOs 
board premiums, equity splits, 198 
Boeker, Warren, 275-76 
Bohan, John, 150 
bonuses, 236, 237, 443n24 

boundary setting and financial capital, 356, 449n26 
Bowerman, Bill, 163 

bricolage framework, hiring decisions, 441n3 
bridge CEOs, 324 
Brin, Sergey, 13 In 

Brody, Mike, 302, 306, 312, 316, 324-26, 353 
Bruderl, J., 430n23, 431n38, 434n4 
bureaucracy blueprint, hiring decisions, 210n, 211 
Burman, Tracy, 156 
bum rates, 43, 252 
Burows, Ken: overview, 24 

commitment shift, 166 

entrepreneurship decision, 51 

equity splits, 154-55, 172, 177 

title designation, 121 
Burton, M. D., 442nl3, 447nl0 
Busenitz, L. W., 425n9, 428n7 
business logic, equity theory, 194-95 
Bussgang, Jeff, 84, 115n, 311, 328n 
Butler, Tim, 14n, 30, 32, 44 

capital See financial capital; human capital entries; social capital 

career dilemma, overview, 8, 27, 29, 59-62. See also entrepreneurship decision entries 

career handcuffs, entrepreneurship decision, 49-50, 66, 431n38 

CareerLeader database, 14, 32, 428n9 

Carland, J. W.,6n 

cash compensation, 162, 177-82, 232-40, 244—45, 442n20. See also equity entries 

cash-only deals, acquisitions, 381-83 

Center for Venture Research, 427n31 

CEO premiums, equity splits, 197-98 

change challenge. Three Rs equilibrium, 200 

Chase, Robin, 54-55, 167, 228 

Chen, Humphrey, 27-28, 31, 37-38, 48, 51-52, 64-65 

Chieh, at govWorks.com, 167-68 

childhood/family background, in entrepreneurship decision, 30-31, 428n9 
Christensen, C. M., 432n50 
Circles, 9, 104-5, 17^77 

Cirne, Lew: entrepreneurship decision factors, 297-98 

exit decision-making, 380-81 

financing decisions, 258, 298, 303-4 

hiring decisions, 232, 345 

motivation understanding, 347, 354 

replacement search, 311-12, 313-14, 315, 322-23 

retention challenges, 319-20, 323-24, 325 

role dilemmas, 318, 319 

serial entrepreneurship benefits, 354 

solo approach decision, 80-81, 298 

succession dilemma, 298-99, 301 

succession/retention challenges, 323-24, 327, 348-49 

succession triggers, 304-5, 308 
classmates, team approach, 104—5, 115-16 



C-level positions: and board interactions, 276-78 
on board of directors, 139-40 

compensation patterns, 233-36, 238-42, 244-45, 442n20 
creating, 218-19, 223-24 
title assignment factors, 121-24, 131-32, 142-43 
top-heaviness prevalence, 119-21 

as upgrading obstacle, 223-24. See also hiring decision entries 
co-CEO arrangements, 131-32, 134-35 
collaborative style, as team approach factor, 80 
commitment blueprint, for hiring decisions, 210n, 211, 231-32 
commitment factor: and equity splits, 156, 157 

in relational teams, 102-3, 190, 436nl9 

as title designation factor, 121 
CommonAngels, 263, 265 

communication challenge. Three Rs equilibrium, 200 

company size, work experience: as entrepreneurship decision factor, 44—46, 430nn29-30 

in hiring decisions, 227-29 
compatibility assessment, 96-99, 113, 142, 350 
compensation committees, board of directors, 289-90 
competitive environment: in idea evaluation, 56-57, 432nn52-53 

as team approach factor, 78, 82, 434nl0 
Computer Associates, 360, 380 
confirmation bias, defined, 59n 
conflict avoidance, 10, 106-11, 142, 200, 436n22 
ConneXus, 27-28, 31, 37-38, 48, 51-52, 64-65 
Connors, Tim, 104, 131, 136, 302 

consensus model, decision-making approach, 129-38, 190-91, 437nnl0, 13-14, 438nnl6-17 
contract law, verbal equity splits, 439nl7 
contributions criteria, equity splits, 150-51, 155-56 
control motivations: and entrepreneurial ideal, 354—61 

and entrepreneurship decision, 32-36 

and equity splits, 162-63, 198, 338-39 

and exit decisions, 377, 380-81, 386 

in financing decisions, 253, 291-95, 447n53 

and hiring decisions, 209-10, 231 

and private equity premium puzzle, 361-63 

and purposeful decision-making, 12-15, 16-17 

in role dilemmas, 141-42, 144 

and solo approach likelihood, 75-76, 434n6 

and succession outcomes, 308-9, 321 

and team size, 81 

control versus wealth decisions, overview: alternative strategies, 343-46, 372-73 
consistency benefits, 346-51 
as dilemma theme, 12-19, 333-43 
future research opportunities, 363-70 
priority shifts, 352-54 

as resource constraints problem, 332-33. See also specific topics, e.g., equity splits; financing approaches; Williams, Evan 
Conway, Ron, 264n 
Cook, Scott, 257, 261 
Cooper, A. C.,432n56 
corporate investors, defined, 445n30 
corporate partnerships, 367-68, 450n51 
Costolo, Dick: overview, 24-25 

acquisition offers, 377-78, 379, 380 

board interactions, 285, 287, 288 

compensation decisions, 236, 240, 244 

entrepreneurial decision factors, 33, 47 

exit decision-making, 377-78, 381, 382-83, 385, 386 



experience factor, 33 
financing decisions, 266-67, 295 
hiring decisions, 216, 220, 221, 225-27, 247-48 
location decision, 369 
motivations factor, 33, 141 
role assignments, 127-28, 139, 141, 192-93 
serial entrepreneurship benefits, 354—55 
in succession process, 342-43 
team formation decisions, 81, 100 
couple -preneur. See Thiers, Genevieve 

coworkers, team approach, 103-4, 115-16, 190-91, 196, 436n20 
creativity, heterogeneous team benefit, 94, 102n, 435nn7-9 
Crews, Bobby, 66, 267 

Crimson Solutions, 151-52. See also Truong, Phuc 
cultural influences, career decision, 30-31, 428n9 
Curtis, Doug, 137, 139 

Dauchy, C. E.,277 

Daugherty, Bill, 132 

David, in Smartix negotiations, 148 

Davoli, Bob, 327 

debt financing option, 253, 443n4 

decision-making control, venture capital impact, 284, 298-99 

decision-making structures, in role dilemmas, 129-38, 190-91, 437nn6, 9-10, 13-14, 438nnl6-17 

Dimon, 194. See also Ludviksson, Georg 

disaster plans, as firewalls, 111-12, 115 

disruptive products, market evaluation, 56, 432n50 

division of labor, as role dilemma, 124-29, 191-93 

DKA, 225-26, 379, 386. See also Costolo, Dick 

Dobrev, S. D., 430n30 

Dorsey, Jack, 342 

earnings puzzle, 15, 18, 361-63, 426n24 
earn-out deals, acquisitions, 381-83 

education factor, in entrepreneurship decision, 37-38, 429nl6 

egalitarian model, decision-making approach, 129-38, 190-91, 437nnl0, 13-14, 438nnl6-17 
Eisenhardt, Kathleen, 132-33, 136, 437n8, 438nl6 
elephant problem, 10, 106-11, 142, 200,436n22 
Elfenbein, D. W., 431nn40, 44 

employer-related changes, entrepreneurship decision, 52-54, 431nn43-44 
engineering blueprint, hiring decisions, 210n 

entrepreneurial experience: as entrepreneurship decision factor, 43-45, 430nn25-26, 438n9 
and equity splits, 156 

as success likelihood factor, 360-61, 449n40 

and title assignments, 122, 124 
entrepreneurial ideal, probability factors, 356-61 
entrepreneurship decision, influences: age patterns, 27-29, 427nn3-5 

childhood/family background, 30-31, 428n9 

idea evaluation, 55-60, 63, 432nn47, 50, 52-53, 433nn58, 60-61 

personal motivations, 32-36, 428n6 

risk-taking preferences, 30, 50, 428n7 

social capital levels, 47-48, 63-64, 95, 431n36, 435nll 
entrepreneurship decision, timing factors: overview, 36-38, 48-49, 59-62, 428nl3, 429nnl4-15 

career handcuffs, 49-50, 66, 431n38 

company size experience, 44—46, 430nn29-30 

education background, 38, 429nl6 

employer-related changes, 52-54, 431nn43-44 

entrepreneurial experience, 43-45, 430nn25-26 



family considerations, 50-52, 54-55, 64-66, 431nn40, 42 
financial capital levels, 48, 54, 430n33, 432n45 
functional background, 39-40, 429n20 
generalist experiences, 50, 431n39 
industry knowledge, 40-43, 429n21, 430nn22-23 
managerial experience, 38-40, 63-64, 429nnl7-18 
nontraditional experiences, 46-47 
equity splits: overview, 145-47, 182-85 

and control motivations, 162-63, 198, 338-39 

criteria for, 150-57, 158, 438nn7-9, 439nl2 

and entrepreneurship experience, 360, 449n40 

equal shares approach, 157, 159-63, 193-96, 439nl3 

formal versus informal agreements, 161-62, 163-65 

and heterogeneity-based conflict, 92-93 

in hiring decisions, 240-45 

negotiation time effects, 161-62, 163-65 

prevalence of, 438n4 

timing decision, 147-50 

venture capitalist impact, 278-80, 283-84. See also Three Rs equilibrium 
equity splits, dynamic agreements: contingency provisions, 171-74 

reasons for, 165-68, 196-97, 440nl9 

self-imposed vesting, 174-77 

template for, 168-71 
equity theory, 194-95 
ethnic comparisons, 91, 434n6 
Evans, Clark, 137,312, 324 
executive derailment, 362n 
exit decisions: overview, 374, 376 

IPO pathway, 376-78, 383-86 

sale route, 376-83. See also succession process, founder-CEOs 
experience factors: and board of directors types, 275-76 

in entrepreneurship decision, 37-40, 43-47, 63-64, 429nnl7-18, 430nn25-26, 29-30 

and equity splits, 156, 160, 360, 449n40 

in hiring decisions, 227-32, 442nl8 

mentor-protege relationships, 445n38 

and solo approach likelihood, 75, 433n2 

and title assignments, 122, 124 
exploitation strategy, likelihood with heterogeneous teams, 94, 435n9 
exploration strategy, likelihood with heterogeneous teams, 94, 435n9 

failure rates, contributing factors, 3-4, 425nnl-2 

family background, in entrepreneurship decision, 30-31, 428n9 

family members, in entrepreneurship decision, 50-52, 54-55, 64-66, 431nn40, 42 

family members, as relational team: hazards of, 100-10, 1 13-14, 436nnl9, 22 

risk reduction strategies, 110-12, 114-15 
FeedBumer. See Costolo, Dick 
financial capital: angel investor statistics, 265-66 

and control versus wealth motivations, 355-56 

and entrepreneurial ideal, 358 

entrepreneurship decision, 48, 54, 430n33, 432n45 

and equity splits, 153-54, 438n8, 439nl3 

as team approach factor, 77-79 

and title assignments, 121, 124 

venture capitalist statistics, 269-71 
financial gain. See wealth entries 
financing approaches: overview, 249-51, 293-95 

angel investors, 262-67 

average requirements, 443n5 



and entrepreneurship timing, 66 

and equity split decision-making, 161-62, 163, 168, 194 

and founder backgrounds, 45 

friends/family, 257-62, 454nl0 

and governance changes, 254, 266-67, 443n6 

as hiring source, 272 

and motivation consistencies, 350 

nonprofit organizations, 364—65, 450n47 

outsider types summarized, 253-56, 257 

and risk anticipation, 359 

self-funding option, 252-53, 255, 256n8 

solo approach correlations, 340 

stages of, 255, 453n7 

as succession trigger, 307-8 

trade-offs summarized, 291-94. See also venture capitalists 
Finkelstein, S., 18n 

firewalls, relational teams, 111-12, 114-15 

Fiske, S. T.,428nl3 

flexibility skills, hiring decisions, 224—29 

Fog Creek Software, 253n 

forgotten founders, equity splits, 148n 

founder decisions, model of, 374, 375 

founders, defining, 6-7, 83-85 

founding dilemmas, overview: commonalities, 9-12, 331-40 
companies listed, 403-10 
company descriptions, 21-25 
definitions, 6-7 
founders listed, 413-22 
future research opportunities, 363-70 
methodology, 19-20, 391-402 
power assumption puzzle, 15, 18-19, 427n4 
private equity puzzle, 15, 18, 361-63, 427n4 
purposeful decision-making benefits, 12-15, 16-17 

types of, 7-9. See also specific topics, e.g., equity splits; team approach decision; Williams, Evan 
France, entrepreneurship motivations, 450n49 
franchising strategy, 356, 368-69, 450n51 
free riders, 148, 177, 440n24 

friends, as relational team: hazards of, 100-10, 113-14, 436nnl9, 22 

risk reduction strategies, 110-12, 114-15 
functional background: and board of directors, 275-76 

as entrepreneurship decision factor, 39-40, 429n20 

and founder-CEO replacement, 315, 447nl0 

and underperformance problem, 223, 442nl3 
future contributions, as equity split criteria, 155-56 
Gates, Bill, 35, 166 

gender comparisons, 32, 33-35, 237-40, 434n6, 443n22 

generalist experiences, as entrepreneurship decision factor, 50, 431n39 

generalists versus specialists, hiring decisions, 224—29, 442nl4 

Glaser, Wffl, 22, 96, 126, 133, 353 

Glass, Noah, 90, 118, 141, 146-47 

Global Entrepreneurship Monitor project, 365-66, 450n49 

golden handcuffs. See handcuffs, in entrepreneurship timing question; vesting schedules 
Gompers, Paul, 290-91, 431n43, 447n51, 449n40, 453n7 
Google: acquisitions by, 118, 377, 381, 382-83, 386 

financing decisions, 250, 264n 

title assignments, 13 In 
govWorks.com, 167-68 

gray-area obstacles, entrepreneurship decision, 61-67 



Greenfield, Jerry, Bin 
Greylock Venture Partners, 303, 308, 315 
growth rate, as control decision, 309, 355 
Quo, Bei, 107-8 

Haleblian, J., 437n7 
Hall, R. E., 370n, 376n 
Hambrick,D. C.,429n20 
Hamilton, B. H.,426n24 

handcuffs, in entrepreneurship timing question, 49-55. See also vesting schedules 

handshake approach, equity splits, 161-62, 163-65, 167 

Hayward, M., 426nl0, 433n61 

Hellman, Thomas, 122, 124, 156, 160, 161, 446n48 

heterogeneous teams, 93-99, 112-13, 160, 435nn7-ll 

Hewlett-Packard, 13 In 

hierarchical model, decision-making approach, 129-30, 134-38, 190-91, 197, 437nn6, 9 
high-potential startups, defmed, 6 

hiring decisions: overview, 8, 205-11, 245-48, 441n2 (ch 8) 

cultural integration challenges, 214—16 

hybrid strategy, 343-44 

and motivation consistencies, 340, 350-51 

relationship framework, 212-14, 216-17, 247, 442n7 

as succession trigger, 305-6 
hiring decisions, rewards framework: overview, 232-33, 247-48 

cash compensation, 233-40, 244-45, 442n20 

equity packages, 240-45 
hiring decisions, role framework: overview, 217-18 

company formalization pressures, 221, 226-27, 442nl5 

company size backgrounds, 227-29 

generalist versus specialist trade-offs, 224-29, 442nl4 

position creation, 218-21 

position upgrades, 222-24 

under-performance problem, 222-24, 442nl3 

work experience considerations, 227-32, 442nl8 
Holodnak, Bill, 218-19, 244 
homogeneous teams: overview, 90-91, 98, 112-13 

benefits, 91-93, 435nn2-3, 6 

compatibility assessment, 96-99 

and equity splits, 160 

and founder-CEO succession, 314-15 

risks, 93-96, 435nn7-ll 
homophily, defined, 90 

Hourihan, Meg: decision misalignments summarized, 186 

equity splits, 145-47 

financing decisions, 249, 257 

as relationship dilemma example, 90, 106 

role conflicts, 117-18, 141 
human capital: from angel investors, 263-64 

and entrepreneurial ideal, 357 

as entrepreneurship timing factor, 37 

as team approach factor, 77-79 

and title assignments, 121, 124 

from venture capitalists, 272-78, 445nn32-33 
human capital, building for entrepreneurship: company size experience, 44—46, 430nn29-30 

education background, 38, 429nl6 

entrepreneurial experience, 43-44, 430nn25-26 

functional background, 39-40, 429n20 

industry knowledge, 40-43, 429n21, 430nn22-23 



managerial experience, 38-40, 429nnl7-18 
non-traditional experiences, 46-47 

schemata development process, 37, 428nl3, 429nl4, 430n22 
Hummingbird, 298 

hybrid strategies, 127-28, 136, 343-46, 372-73, 438nl6, 448nl9 

idea evaluation, pre-founding stage, 55-58, 432nn47, 50, 52-53, 433nn58, 60-61 

idea people and title designations, 121-23 

idea premium, equity splits, 151-53, 197-98, 438n7 

ideal types, analytical purposes, 436n5 

industry knowledge factor: entrepreneurship decision, 40-43, 429n21, 430nn22-23 

solo approach decision, 434n4 
inertia challenge. Three Rs equilibrium, 201-3 
instinct problem, as dilemma commonality, 10-11 
international comparisons, 365-66 
Intuit, 257, 261 

investor dilemmas, overview, 8, 205-8, 249-51. See also financial capital; financing approaches; venture capitalists 
IPO decision, 37^78, 383-86, 449n40 
IPO statistics, in methodology, 20, 427n30 
iWon, 132 

jack of all trades. See generalist entries 
Jobs, Steve: equity splits, 152 

friendship obstacles, 103, 108-9 

labor division, 124 

on rational decision-making, 60 

team values misalignment, 97 
Joint Juice, 75 
jungle metaphor, 84 

Kaleil, at govWorks.com, 167-68 
Kamen, Dean, 320, 346, 351, 356 
Kauffman Foundation, 14 
Kawasaki, Guy, 59 
Kendle, Candace, 383-84 
KhuUer, Vrvek: overview, 23 

equity splits, 148, 155 

role assignments, 125-26, 192 

team formation decision, 79-80, 95 
KhuUer, Vrvek, entrepreneurship decision factors: age and risk, 50 

education background, 38, 43-44 

family considerations, 51, 65 

financial capital levels, 48 

functional background, 41 

idea evaluation, 55, 56 

social capital levels, 47-48 
King outcome, overview, 13-14, 337-40. See also corAroX entries 
Knight, Phil, 163 

Kraft, Jon: company overview, 21-22 

decision-making structure, 133, 353 

entrepreneurship decision factors, 64 

role assignments, 126, 129 

team formation decisions, 95-96 
Kraus, Janet, 9, 104-5 

L90, 150, 383. See also Addante, Frank 

labor assignments, as role dilemma, 124—29, 191-93 

layoff possibility, as hiring consideration, 217 



Lazaridis, Michael, 132n 
Lerner, Josh, 277-78, 453n7 

liquidation preferences, 250n, 280-82, 378, 446nn44-45 
location question, 51, 368-69 
LovaUo, D., 433n60 

Ludviksson, Georg: equity splits, 147, 169-71, 173-74, 194, 201-2 

heterogeneity-based conflict, 92-93 

team formation decisions, 84—85 
Lunt, Eric, 127-28, 354-55, 369 
Lynx Solutions. See Milmo, James 

Malhotra, Deepak, 171 

managerial experience: as entrepreneurship decision factor, 38-39, 63-64, 429nnl7-18 

and solo approach likelihood, 75, 433n2 
market potential questions, in idea evaluation, 56, 432n50 
Marx, Matt, 101, 103, 195, 284 
Masergy. See NaUs, Barry 
Mayfield, 168 
McCall, Morgan, 362n 
McCance, Henry, 308, 315, 327 
McManus, Paul, 100, 115 
McQuillen, Tim, 191 
Megaserver, 306,316, 325 
Meisenheimer, Mike: overview, 24 

board role dilemma, 140 

cash compensation, 178 

decision alignments summarized, 187-88 

decision-making approach, 134—35 

entrepreneurship decision factors, 40, 66 

equity splits, 153, 177 

financing decisions, 252 

role conflicts, 140, 190-91 
mental models. See human capital, building for entrepreneurship 
mentor-protege relationships, experience factor, 445n38 
methodology, overview, 19-21, 391-402 
Meyer, G. D., 429nl8 
Microsoft, 166 

milestone -based vesting, 175-77 
Mills, C. Wright, 18-19 

Milmo, James: board interactions, 139-40, 286, 289 
on CEO replacement, 312 
decision-making structure, 136-37 
entrepreneurship decision, 63 
equity splits, 152 

exit decision-making, 282, 352-53, 378, 379, 382, 383 

financing decisions, 282, 293, 378 

hiring decisions, 214, 224, 227, 231 

salary structure, 179-80, 181-82 

on succession process, 324 

title assignments, 137 
Mittal, Saurabh, 80, 155 
Mobilink, 382, 383 
Monarch Capital Partners, 252 

Mr. Right versus Mr. Right Now, 229-30. See also experience factors 

Nails, Barry: overview, 23, 357-58 
board interactions, 39, 45-46, 288 
financing decisions, 252-53 



hiring decisions, 223-24, 228, 229 

location decision, 368-69 

solo approach decision, 74-75 
Nails, Barry, entrepreneurship decision factors: age, 28 

childhood/family background, 3 1 

company size experience, 45-46 

employer-related changes, 53 

family obstacles, 51, 368-69 

idea evaluation, 55-56, 63 

industry knowledge, 41 

managerial experience, 38-40 

social capital levels, 47 

timing regret, 66 
Nanda, R.,454nl0 

natural biases problem, as dilemma commonality, 10-11 

Nazeeri, Furqan, 288 

networks. See social capital 

New Relic, 360, 380-81. See also Cirne, Lew 

Nike, equity splits, 163 

noncompete agreements, 49-50, 103n 

nonprofit comparisons, 364—65, 450n47 

nontraditional experience factor, entrepreneurship timing, 46-47 
Noro-Moseley, 292, 293 
Nyman, Vic, 311 

Odeo. See Williams, Evan 
Olechowski, Steve: exit decision-making, 381 
location decision, 369 

role assignments, 127-28, 139, 141, 192-93 

serial entrepreneurship benefits, 354—55 

team formation decisions, 81 
opportunity costs, as equity split criteria, 154-55 
opportunity questions, in idea evaluation, 57-58 
optimism/overconfidence: as dilemma commonality, 10-11, 338n, 362 

and equity splits, 157, 165-66 

in idea evaluation, 58-59, 433nn58, 60-61 

and liquidation preferences, 281-82, 294 

prevalence of, 425nn8-9 

and resource attraction, 426nnl0-ll 

and title conflicts, 121 
Ostrander, S. A., 450n47 
outsourcing strategy, 356, 449n26 
overlapping roles, divided labor compared, 124-27 

Packard, David, 13 In 
Page, Larry, 13 In 

Pandora Radio. See Westergren, Tim 
Panel Study of Entrepreneurial Dynamics, 14, 78, 434n5 
participation element, liquidation preferences, 28 In, 446n44 
Pascal, Javier: board interactions, 139-40, 286-87 

decision-making structure, 136-38 

entrepreneurship decision, 63 

equity splits, 152 

exit decision-making, 282, 378, 382 
financing decisions, 282, 294, 378 
salary structure, 179-80, 181-82 
passion factor: entrepreneurship decision, 8, 10-11, 30, 428n6 
in idea evaluation, 29, 58-60 



past contributions, equity split criteria, 150-51 

path dependence problem, dilemma decisions, 10 

PayPal, financing decisions, 264n 

Pennings, J. M.,432n52 

planning fallacy, 433n60 

Playing-with-Fire Gap, 109-12, 115, 21^17 

power assumption puzzle, 15, 18-19, 361-63, 427n4 

pre-founding stage. See career dilemma, overview; entrepreneurship decision entries 
private-equity premium, puzzle of, 15, 18, 361-63, 426n24 
product development success, as succession trigger, 304-7 
professional services firms, C-level position arrangements, 436n2 
Proteus Biomedical, 57 

psychological needs, as team approach factor, 80-81 

quick-equal teams, equity splits, 161-62, 163-65 

Ratner, Dan, 23-24, 54, 110-12, 309 
ReaXions, 150, 162 

Reich, Michael: equity spUts, 147, 152, 168-71, 173, 197, 201-2 

heterogeneity-based conflict, 92-93 

team formation decisions, 81, 84-85 
relational teams: hazards of, 100-10, 113-14, 436nnl9, 22 

reward implications, 193-94 

risk reduction strategies, 110-12, 114-15 

role implications, 189-90 
relationship dilemmas: overview, 8, 89-90, 112-16 

financing decisions, 258-62, 265 

in founder transition process, 324—25 

hiring decisions, 212-15, 216-17, 245-48, 442n7. See also team approach decision; Three Rs equilibrium; Williams, Evan 
Renaissance Capital, 427n30 
Research in Motion, Ltd., 132n 

resource constraints, as dilemma foundation, 331-33. See also control entries 
resource dependence, 13, 332-337, 441n2 (Part 111 intro) 
Ressi Adeo, 294n 

rewards dilemmas, overview, 8. See also cash compensation; equity splits; Three Rs equilibrium; wealth entries 

Rich outcome, trade-offs summarized, 13-14, 337-40. See also wealth entries 

risk anticipation and entrepreneurial ideal, 359 

risk aversion factor: entrepreneurship decision, 30, 50, 428n7 

and equity splits, 157 

and exit decisions, 385 
risk reduction strategies, team approach decision, 110-12, 113-15 
Rochefoucauld, Frangois de La, 60 
Rockefeller, John, 104 
role dilemmas: overview, 8, 117-19, 142-44 

and control motivations, 141 — 42 

decision-making structures, 129-38, 437nn6, 9-10, 13-14, 438nnl6-17 
founders on boards, 138-40 
labor assignments, 124—29 
in succession process, 325-26 
title assignments, 119-24 

and wealth motivations, 141-42. See also Three Rs equilibrium 
role dilemmas, hiring decisions: overview, 217-18, 245-48 

company formalization pressures, 221 

generalist versus specialist trade-offs, 224-29, 442nl4 

position creation, 218-21 

position upgrades, 222-24 

underperformance problem, 222-24, 442nl3 
Rubbish Boys, 80, 253, 259, 356 



Ruef, Martin, 193, 434n6 



Sachleben, Mark, 232 
Sahlman, Bffl, 252, 291 

salary structures, 178-82, 230-31, 233-34, 237-40, 442ii20 
Savage, George, 57 

Savage Beast, name change, 353. See also Westergren, Tim 

schemata development, 37, 428nl3, 429nl4, 430n22 

Schilling, Curt, 35, 42-43 

Schnoor, Bill, 103n 

Schumpeter, Joseph, 6n 

Scudamore, Brian, 80, 356 

Segway, 320, 346, 351,356 

self-funding option, 252-53, 255, 256n8 

self-initiated succession, founder-CEOs, 301-2, 318, 328 

Sequoia, 271-72 

serial entrepreneurs, research opportunities, 360. See also entrepreneurial experience 

serial premiums, equity splits, 156 

724 Solutions, 382 

Shane, Scott, 38, 41, 166 

Sherbrooke, Kathy, 9, 104-5 

Shobe, Matt, 127-28 

Sigma Partners, 327 

Sittercity. See Thiers, Genevieve 

slow-equal teams, equity splits, 161-62, 163-65 

slower-growth rate, as control decision, 309, 355 

Smartix. See KhuUer, Vrvek 

social capital: and angel investors, 263-64 

as control leverage, 358, 449n33 

and entrepreneurial ideal, 357-58 

as entrepreneurship decision factor, 47-48, 63-64, 95, 431n36, 435nll 

and financial capital attraction, 48, 357-58, 431n35 

for hiring decisions, 212-15, 272 

and homogeneous teams, 94-95, 435nl0 

as team approach factor, 77-79 

and title assignments, 121, 124 

fi-om venture capitalists, 271-72, 445n28 
social logic, equity theory, 194-95 
solo approach decision: arguments against, 76-81 

arguments for, 73-76, 433n2 

dilemma overview, 8, 83, 85-87 

environmental context, 82-83 

statistics, 73, 74, 78, 434nl4 

wealth trade-off, 339-40 
Sonic Mountain, 342 

specialist experience: as entrepreneurship obstacle, 50 

hiring decisions, 224—29 
split board votes, 287-88 
Spolsky, Joel, 253n 
Spotlight, 382 
Spyonit, 216, 220, 382 

staging strategy, board control, 290-91, 447n51 
Stam, E., 31 

Stanford Project on Emerging Companies (SPEC), 210-11, 440nl9, 441 nl (Part III intro) 
"star" blueprint, hiring decisions, 210n, 231-32 
Steinman, Jonas, 132 
Steria, 349 

Stevenson, Howard, 6-7 



stewardship theory, 180-82, 363 
Stinchcombe, Arthur, 3, 74, 81, 434nl3 
Stone, Kevin, 75 

Strohm, David, 311, 314, 318, 327 

StrongMaiL See Addante, Frank 

structural leverage metric, hiring decisions, 220-21 

success-based triggers, succession process, 303-9 

succession process, founder-CEOs: overview, 8, 299-301, 326-28 

board-initiated transition, 302-3 

and motivation consistencies, 351 

and power assumption puzzle, 18-19 

replacement search, 310-15, 447n9 

retention trade-offs, 317-21 

self-initiated transition, 301-2, 318, 320 

success-based triggers, 303-9 

transition challenges, 316-26 
supermajority rights, venture capitalists, 286 
support/validation needs, as team approach factor, 80 
sustaining products, market evaluation, 56, 432n50 
Switzerland, entrepreneurship motivations, 450n49 

Talmudic wisdom, 288n 

task preferences, as team approach factor, 79 

taxes, 149n, 165n 

team approach decision: overview, 8, 89-90, 112-16 

arguments against, 73-76, 433n2 

arguments for, 76-81 

classmate opportunities, 104-5 

compatibility assessment, 96-99, 350 

dilemmas summarized, 8, 69-71, 83, 85-88, 189-93 

diversity benefits, 93-96, 112-14, 435nn7-ll 

environmental context, 82-83, 434nl0 

homogeneity benefits, 90-93, 435nn2-3, 6 

previous-coworker advantages, 103-4, 436n20 

relational team hazards, 100-10, 436nnl9, 22 

risk reduction strategies, 110-12, 113-15 

size considerations, 81-82 

statistics, 73, 74, 78, 434nl4 

title decisions, 83-85 
TechCoast Angels, 263, 265 
Tedlow, Richard, 104 
TheFunded.com, 294n, 359n 
Thiers, Genevieve: overview, 23-24, 344—45 

entrepreneurial decision factors, 33, 53-54, 55 

financing decisions, 345 

hiring decisions, 212, 214, 231-32 

location decision, 369 

team formation decisions, 110-12 
38 Studios, 35, 42-43 
Thompson, Andrew, 57 

Three Rs equilibrium: challenges to, 186-89, 199-203 

relationship-reward linkages, 193-97 

relationship-role linkages, 189-93 

role-reward linkages, 197-99 
time -based vesting, 175 
title assignments: overview, 117-18, 142-43 

correlational factors, 121-24 

and decision-making structures, 129, 131-32, 137 



as position upgrading obstacle, 223-24 

top-heaviness arrangements, 119-21 
Tom, at govWorks.com, 167-68 
Trachtman, Les: exit decision-making, 379 

in founder-CEO succession process, 302, 305-6, 312, 316, 322, 324-26 
Transcentive. See Trachtman, Les 
Triandiflou, Jim: overview, 24 

board interactions, 140, 288-89, 295 

cash compensation, 178 

decision aUgmnents summarized, 187-88 

decision-making approach, 134-35 

entrepreneurship decision factors, 40, 49, 66 

equity splits, 154, 173, 177 

financing decisions, 252, 267, 292, 295, 356 

role conflicts, 121, 190-91 

team formation decision, 79, 103 

title designation, 121 
Truong, Phuc: cash compensation, 178 

equity spUts, 151-52, 155, 169-71,201-2 

heterogeneity-based conflict, 92-93 

team formation decision, 81, 84—85 
turnover rates, correlations, 101, 103, 284, 442nl3, 446n46, 447nl0 
Twitter, 4, 342-43. See also Williams, Evan 

uncertainty planning, dynamic equity agreements, 171-74 
UpDown. See Reich, Michael 
value systems, team compatibility assessment, 96-99 
The Venture Capital Cycle (Gompers & Lemer), 453n7 
venture capitalists: overview, 267-69 

angel investments compared, 263, 265-66 

benefits from, 269-78, 445nn28, 32-33 

and board operation changes, 285-91, 446nn48-50, 447n51 

company exit statistics, 268-69, 444n22 

and decision-making control changes, 284—85, 298-99 

and egalitarian teams, 437nl3 

and exit scenarios, 280-83, 286, 446nn44-45 

IPO benefits, 384-85 

and ownership changes, 278-80, 283-84, 446n46 
and salary structures, 179-80, 181 
tradeoffs summarized, 291-93 

valuation costs, 272, 445n35. See also angel investors; financing approaches; succession process, founder-CEOs 
Venture Expert Database, 427n31 
vesting, defined, 146n 

vesting schedules, 152, 165n, 174-77, 242-44, 283 
veto rights, venture capitalists, 286 
voluntary succession, founder-CEOs, 301-2, 318, 320 
VP -level positions, 218, 233-34, 236, 238-42, 318-19 

Waldroop, James, 14n 

Warren, with UpDown, 84-85, 168, 169-71 

wealth motivations: and entrepreneurial ideal, 354—56 

and entrepreneurship decision, 32-36 

and equity splits, 157, 198 

and exit decisions, 377-78, 386 

in financing decisions, 291-94, 447n53 

and hiring decisions, 210 

and purposeful decision-making, 12-15, 16-17 

in role dilemmas, 141-42, 144 



and succession processes, 302, 321-22 
and team values misalignment, 97 
of venture capitalists, 268 
wealth versus control decisions, overview: alternative strategies, 343-46, 371-73 
consistency benefits, 346-51 
as dilemma theme, 12-19, 333-43 
future research opportunities, 363-70 
priority shifts, 352-54 

as resource constraints problem, 332-33. See also specific topics, e.g., Costolo, Dick; equity splits; financing approaches 
Weber, Max, 436n5, 437n6 
Wellie, with Crimson Solutions, 152 
Wellman, David, 66-67 
Westergren, Tim: company overview, 21-22 

decision-making structure, 133, 353 

entrepreneurship decision factors, 46-47, 52, 55, 63-64 

equity splits, 160 

financing decisions, 257, 261 

hiring decisions, 212-13, 217, 231 

role assignments, 126-27, 128-29, 192 

team formation decision, 78, 95-96 
Williams, Evan: acquisition offers, 377, 378, 381, 382 

board interactions, 276, 277, 285-86, 348 

company overview, 3-4 

decision consistencies/misalignments summarized, 186, 340-43, 348, 352, 353-54 

dilemmas summarized, 4—5 

entrepreneurial motivations, 32-33, 348, 352 

equity splits, 145-47, 197 

exit decision-making, 377, 378 

financing decisions, 249-50, 282 

hiring decisions, 209-10, 211, 217 

role conflicts, 117-18, 141 

team formation decisions, 89-90, 106 
Williams, Richard, 299, 312, 315, 318, 322-23, 325 
Wily Technology. See Cirne, Lew 
work experience. See experience factors 
Wozniak, Steve, 97, 103, 108-9, 124, 152 



Zipcar, 54-55, 167, 228 

Zondigo, 119, 151, 228-29. See also Addante, Frank